A Novel Arabic CAPTCHA Generation Framework Based on Pseudoisochromatic Plates for Enhanced Human Verification

A Novel Arabic CAPTCHA Generation Framework Based on Pseudoisochromatic Plates for Enhanced Human Verification | Research Square window.SnipcartSettings = { analytics: { enabled: false } }; (function() { var accessVector = localStorage.getItem('access_vector') || ''; window.dataLayer = window.dataLayer || []; if (accessVector) { window.dataLayer.push({ user: { profile: { profileInfo: { snid: accessVector } } } }); } })(); (function(w,d,s,l,i){w[l]=w[l]||[];w[l].push({'gtm.start':new Date().getTime(),event:'gtm.js'});var f=d.getElementsByTagName(s)[0],j=d.createElement(s),dl=l!='dataLayer'?'&l='+l:'';j.async=true;j.src='https://www.googletagmanager.com/gtm.js?id='+i+dl;f.parentNode.insertBefore(j,f);})(window,document,'script','dataLayer','GTM-K279D39R'); Browse Preprints In Review Journals COVID-19 Preprints AJE Video Bytes Research Tools Research Promotion AJE Professional Editing AJE Rubriq About Preprint Platform In Review Editorial Policies Our Team Advisory Board Help Center Sign In Submit a Preprint Cite Share Download PDF Article A Novel Arabic CAPTCHA Generation Framework Based on Pseudoisochromatic Plates for Enhanced Human Verification Suliman A. Alsuhibany This is a preprint; it has not been peer reviewed by a journal. https://doi.org/ 10.21203/rs.3.rs-7367912/v1 This work is licensed under a CC BY 4.0 License Status: Posted Version 1 posted You are reading this latest preprint version Abstract This paper presents a novel Arabic CAPTCHA generation framework that leverages Pseudoisochromatic Plates (PIPs)—a technique traditionally used in color vision testing—to enhance security and usability in human verification systems. The proposed scheme embeds both printed and handwritten Arabic texts into PIP-style backgrounds using randomized color blending and non-font-based character rendering. This design introduces high visual complexity that resists segmentation and recognition by automated solvers while maintaining legibility for human users. A comprehensive experimental evaluation involving human participants and machine recognition tools demonstrates the robustness of the proposed approach against state-of-the-art attacks, including Google Vision API. Usability assessments further indicate that, while handwritten text embedded in PIP backgrounds achieves superior security, printed text offers better efficiency and effectiveness. The results underscore a critical trade-off between CAPTCHA security and usability and highlight the potential of PIP-based mechanisms to improve Arabic CAPTCHA systems. This work introduces a new direction in CAPTCHA design tailored for linguistically complex scripts and provides valuable insights for developing secure and user-inclusive verification tools in Arabic-speaking digital environments. Physical sciences/Engineering Physical sciences/Mathematics and computing Security Arabic CAPTCHA Usability Pseudoisochromatic Plates Experimental study AI-Resistant CAPTCHA Figures Figure 1 Figure 2 Figure 3 Figure 4 Figure 5 Figure 6 Figure 7 Figure 8 Figure 9 Figure 10 1. Introduction The widespread reliance on online services has intensified the need for secure and user-friendly human verification mechanisms. Completely Automated Public Turing tests to tell Computers and Humans Apart (CAPTCHAs) are widely deployed to prevent automated bots from abusing web services [ 1 ]. While traditional CAPTCHAs, such as distorted text, image selection, or click-based puzzles, have proven effective to some extent, they are increasingly vulnerable to advanced machine learning and sophisticated optical character recognition (OCR) techniques (e.g, [ 2 ]). Moreover, CAPTCHAs should also balance security with usability to avoid frustrating legitimate users [ 3 ]. Arabic, as one of the most widely spoken languages globally and a primary language in many digital markets, presents unique challenges in the context of CAPTCHA design. The cursive nature of Arabic script, its contextual letter shapes, and diacritical marks pose difficulties for both OCR systems and CAPTCHA generators. The groundbreaking papers on Arabic CAPTCHAs were [ 4 – 5 ] as Arabic letters were shown in the CAPTCHAs. These were followed by a set of studies (e.g., [ 6 – 9 ]) that used both Arabic printed and handwritten letters. The output of these studies showed promising results. Despite these results, most existing Arabic CAPTCHA schemes are either direct adaptations of English models or rely on distorted handwritten fonts, making them susceptible to pattern recognition-based attacks or difficult for users to solve. There is a pressing need for CAPTCHA approaches that are inherently more secure, linguistically appropriate, and accessible to a diverse population. Thus, this paper proposes a novel Arabic CAPTCHA generation method based on Pseudoisochromatic Plates (PIPs) —a visual technique originally developed for diagnosing color vision deficiencies [ 10 ]. PIPs consist of color-blended backgrounds embedded with readable characters or shapes discernible only to individuals with normal color perception. By adapting this concept, we introduce a CAPTCHA model that leverages subtle color variations and visual noise to obscure Arabic text in a way that is easily interpreted by humans but yet challenging for automated solvers. Unlike conventional noise-based or distorted-text CAPTCHAs [ 11 ], PIP-inspired CAPTCHAs exploit the complexity of human color perception, thereby offering a new layer of resistance against machine-based attacks. As an illustration of the proposed approach, Fig. 1 presents representative samples generated using our developed generator. Specifically, (a) displays a sample utilizing printed Arabic text, while (b) shows a counterpart rendered in handwritten style. Although recent studies (e.g., [ 11 ][ 12 ]) have emphasized the superior resistance of handwritten Arabic CAPTCHAs to automated attacks, incorporating the printed text style within this framework allows for a comparative investigation into the security and usability trade-offs—thereby contributing further insights to the existing literature. This dual-style integration not only broadens the applicability of the proposed CAPTCHA system across diverse user groups, but also facilitates a comprehensive evaluation of human interaction dynamics and machine-solving resistance. By juxtaposing both styles within a unified generation framework, the proposed method enables a deeper understanding of how visual complexity, font variability, and contextual legibility influence CAPTCHA effectiveness. Furthermore, the use of the PIPs technique enhances visual noise and perceptual ambiguity—essential features for countering automated recognition tools—while maintaining readability for human users. This balance is critical for designing CAPTCHAs that are both user-friendly and resilient against adversarial machine-learning models. The proposed CAPTCHA scheme was evaluated in terms of both usability and security through a series of experimental studies. The results demonstrate that human users were able to solve the CAPTCHA with significantly higher effectiveness compared to machine-based algorithms. Furthermore, the findings indicate strong resistance to automated attacks, highlighting the robustness of the proposed scheme. The proposed approach contributes to the field in several ways. First, it innovatively integrates PIP principles into CAPTCHA design, enhancing security through perceptual obfuscation rather than distortion alone. Second, the approach is evaluated through a series of experiments involving human participants and automated attack models to assess both usability and robustness. Finally, it categorizes exist studies based on the type of text employed. Thus, this study explores a novel direction in CAPTCHA development by fusing Arabic script characteristics with the perceptual challenges of PIPs. The findings aim to advance the design of CAPTCHAs that are both secure and inclusive, particularly in Arabic-speaking and visually diverse user communities. The remainder of this paper is organized as follows: Section 2 reviews the related work. Section 3 provides an overview of the PIPs test. Section 4 outlines the proposed methodology, while Section 5 describes the experimental setup and evaluation. Section 6 presents the results, which are further discussed in Section 7. Finally, Section 8 concludes the paper. 2. Related Work To the best of our knowledge, this paper is the first work for exploiting PIP approach to be applied as text-based Arabic CAPTCHAs. Therefore, this section shows a comprehensive review of the state of the art of Arabic CAPTCHA schemes. Arabic printed text-based CAPTCHAs have been widely explored for enhancing online security. Shirali-Shahreza and Shirali-Shahreza introduced "Baffletext" CAPTCHA for Persian and Arabic scripts, emphasizing text distortion to counter OCR attacks [ 4 ]. Shahreza applied Arabic CAPTCHA specifically to spam SMS verification [ 5 ], while Khan et al. evaluated its effectiveness against automated intrusions [ 6 , 13 ]. Banday and Sheikh reviewed various CAPTCHA types, highlighting Arabic variants [ 14 ]. To increase unpredictability, Sulaiman and Hassan proposed random Arabic character generation [ 15 ], and Alsuhibany et al. analyzed difficulty and robustness against OCR and segmentation attacks [ 16 , 18 ]. Sulaiman also explored chaotic maps to introduce unpredictability [ 17 ]. Moreover, multilingual [ 19 ] and stylistically complex Nastaliq script-based CAPTCHAs were introduced to further enhance security [ 20 ]. Collectively, these studies underscore the importance of linguistic complexity, randomness, and stylistic diversity in strengthening Arabic CAPTCHA systems against automated attacks. Recent research has explored various Arabic CAPTCHA approaches, emphasizing security through handwritten and multilingual methods. Alsuhibany and Parvez [ 1 ] developed a secure Arabic handwritten CAPTCHA resistant to OCR attacks, while Alrasheed and Alsuhibany [ 3 ] employed adversarial techniques to further enhance security. Alsuhibany and Alquraishi [ 21 ] explored Arabic CAPTCHA usability via visual cryptography, and Parvez et al. [ 22 ] introduced an Arabic CAPTCHA gamification framework for cybersecurity education. Lajmi et al. [ 23 ] strengthened security through calligraphy-based handwritten CAPTCHAs, whereas Aldosari [ 24 , 26 ] presented multilingual handwritten CAPTCHA methods. Additionally, Parvez and Alsuhibany [ 25 ] proposed a segmentation-validation technique to generate robust handwritten Arabic CAPTCHAs. Collectively, these studies demonstrate the effectiveness of handwriting styles, multilingualism, and advanced validation techniques in enhancing Arabic CAPTCHA security and usability. Fawa’reh et al. [ 27 ] proposed an Arabic CAPTCHA scheme designed to mitigate deep learning-based attacks targeting text image CAPTCHAs. Their approach introduces unique visual distortions and features tailored to the Arabic script, enhancing resistance to automated recognition methods commonly employed in adversarial machine learning. The study highlights the effectiveness of culturally adapted CAPTCHA designs in improving security against sophisticated AI-driven threats. Recent research has focused on enhancing the interactivity and security of Arabic CAPTCHAs. Alsuhibany and Parvez [ 12 ] proposed attack-filtered interactive Arabic CAPTCHAs that adapt based on detected attack patterns, aiming to increase robustness against automated solvers. In a comparative study, Alsuhibany and Alnoshan [ 28 ] evaluated interactive handwritten and text-based handwritten Arabic CAPTCHA schemes tailored for mobile devices, highlighting trade-offs between usability and resilience. Additionally, Alsuhibany and Alquraishi [ 21 ] employed visual cryptography to improve the security and usability of Arabic text-based CAPTCHAs, offering a novel user-verifiable authentication mechanism. Several studies have explored Arabic reCAPTCHA systems aimed at improving web security and digitizing Arabic content. Bakry et al. [ 8 ] introduced ARECAPTCHA, which leverages crowdsourcing from native speakers to digitize Arabic text. Building on this, Abubaker et al. proposed a cloud-based Arabic reCAPTCHA service, detailing its design and architecture across multiple works [ 9 , 10 ], and emphasizing its application in enhancing the digitization of Arabic manuscripts [ 30 , 31 ]. Their research also addressed the broader challenges and opportunities of Arabic digital content [ 29 ], presenting the reCAPTCHA system as a dual-purpose tool for both security and heritage preservation. To provide a clearer understanding of the current landscape of Arabic CAPTCHA research, the reviewed studies are summarized in Table 1 . This table categorizes each study based on the type of text employed—whether meaningful (semantically valid Arabic words or phrases), meaningless (random or distorted characters), or individual characters. Such classification helps highlight trends in usability, readability, and security across different CAPTCHA designs. Table 1 Summary of Reviewed Arabic CAPTCHA Studies by Text Type (Meaningful vs. Meaningless) Type of Arabic CAPTCHAs Study Meaningless Meaningful Characters Printed text-based [ 4 – 6 ] [ 13 – 15 ][ 17 ] [ 7 ][ 16 ][ 18 ] [ 19 – 20 ] Handwritten text-based [ 24 – 26 ] [ 1 ][ 3 ][ 21 – 22 ] [ 23 ] Image-based - - [ 27 ] Interactive [ 12 ][ 28 ] [ 21 ] - reCAPTCHA - [ 8 – 10 ][ 29 ][ 30 ][ 31 ] 3. Pseudoisochromatic Plate Test: An Overview PIP tests are widely recognized as standard tools for screening, diagnosing, and classifying color vision deficiencies (CVD) [32]. These tests typically consist of colored dot patterns that form identifiable figures—such as numbers, letters, lines, or geometric shapes—that are distinguishable from the background only through color differences. The visibility of these figures depends on chromatic contrasts; for individuals with CVD, such contrast may be imperceptible, rendering the figure invisible or misidentified [33]. The effectiveness of these tests is rooted in the opponent-process theory of color vision, which posits that visual perception is governed by three antagonistic color pairs: black-white, red-green, and blue-yellow [34]. In individuals with color vision deficiency, two or more colors from an opponent pair may appear identical, resulting in isochromatic perception where the figure becomes indistinguishable from its background. The concept of PIP testing originated with Jakob Stilling, who developed the first clinical color vision test using this principle [32]. Although Stilling’s test was groundbreaking, it had several limitations, which were later addressed by the development of the Ishihara plates , now among the most widely used PIP tests worldwide [35]. Another well-known variant is the Hardy-Rand-Rittler (HRR) test, which uses geometric symbols like circles, crosses, and triangles in a specific chromatic arrangement to challenge color-deficient observers [36]. In recent years, digital implementations of PIP tests have emerged to improve standardization and repeatability. Computerized systems, such as the Cambridge Colour Test (CCT) , offer adaptive difficulty and precise control over chromatic parameters. The CCT presents a colored Landholt C that must be identified based solely on chromatic differences, enabling detailed measurement of color discrimination thresholds [37]. Figure 2 illustrates a sample for aforementioned test types. Overall, PIP tests remain indispensable in both clinical and research settings due to their simplicity, non-invasiveness, and diagnostic value. The transition from static printed plates to digital, interactive formats underscores their adaptability to modern needs. This versatility opens new avenues for novel applications—one of which is the generation of visual challenges in CAPTCHA systems. Specifically, the inherent difficulty PIP designs pose for automated systems, while remaining interpretable by humans with normal color vision, makes them particularly well-suited for developing secure, user-friendly Arabic CAPTCHAs. By embedding Arabic text within color-differentiated figure-background patterns, akin to those used in PIP tests, it is possible to create CAPTCHA schemes that are resistant to machine learning-based attacks while maintaining high usability for native Arabic speakers. However, since PIP tests are widely recognized as standard tools for screening, diagnosing, and classifying color vision deficiencies, careful refinement will be necessary to ensure that the proposed approach is accessible to all users—not only those with normal color vision. This consideration will be further discussed in the following section. 4. Methodology This section outlines the design and evaluation of our PIP–based Arabic CAPTCHA system. We first describe the sample generation process for both printed and handwritten text styles using PIPs. We then present our testing procedures—covering adversarial resilience, usability assessments, and evaluation metrics—to demonstrate the framework’s security and user-friendliness. 4.1. Framework of the Proposed CAPTCHA Scheme This section introduces the proposed CAPTCHA design framework, which combines Arabic text generation with a PIP-style background to develop a secure and user-accessible challenge mechanism, as shown in Fig. 3 . The framework follows a structured pipeline that begins with the generation of Arabic text or words, followed by an embedding phase in which the text is integrated into a PIP-like background using a multistage color blending process. This embedding stage consists of three key phases: color selection , word representation , and image rendering —each contributing to the visual complexity and robustness of the CAPTCHA. To ensure comprehensive evaluation, the framework also incorporates two core considerations: security and usability . The security dimension focuses on resistance to current automated recognition techniques and challenges posed by color-blind bots, while the usability dimension emphasizes human readability, accessibility across diverse visual capabilities, and suitability for various user groups. These phases are discussed in the following sections. 4.1.1. Generation of Arabic text/word This subsection outlines the procedures used to generate the Arabic text that forms the core content of the proposed CAPTCHA scheme. Two distinct types of Arabic text generation are considered: printed and handwritten . Each type serves a unique role in assessing the system’s robustness and usability. To facilitate structured implementation and analysis, the generation process is divided into two categories: (1) Printed Arabic text generation , and (2) Handwritten Arabic text generation . Each category employs its own dedicated generation method tailored to its respective style and application context. Accordingly, we employed the text-generation tool described in [ 21 ] to produce both meaningful and meaningless printed Arabic samples. Representative outputs are shown in Fig. 4. Meaningless handwritten samples were produced using the text-generation tool described in [ 38 ], while meaningful handwritten words were generated according to the approach outlined in [ 3 ]. Representative outputs are shown in Fig. 5. 4.1.2. Embedding Arabic text in a PIP-like background using color blending. Once the printed or handwritten Arabic text is generated (cf. Figures 4 and 5), it is embedded within a PIP–style background via a three-phase color‐blending procedure: color selection , word representation , and image rendering . Phase 1: Color Selection In this phase, background and foreground hues are chosen from a predefined palette using a randomized—but constrained—selection process. Randomization is informed by empirical findings reported in [ 39 ], which identified and excluded combinations with poor human performance (e.g., yellow-on‐white: 84% accuracy, 8.5s completion time; white‐on‐yellow: 79% accuracy, 9.8s) to ensure both usability and security aspects. Consequently, only color pairs demonstrating acceptable recognition rates and response times are retained. Phase 2: Word Representation Rather than rendering text with standard fonts, each CAPTCHA word is depicted through an assembly of randomly sized and colored circles. These circles vary in chromaticity and luminance according to the selected color pair, producing unique character shapes in each instance. This stochastic, non-font‐based representation increases the difficulty for automated segmentation and character‐recognition algorithms, while remaining legible to human observers. Phase 3: Image Rendering In the final phase, the colored-dot word representation is composited onto the chosen PIP-style background. The generator merges foreground and background layers, applies anti‐aliasing, and outputs the result as a high‐resolution, lossless image. Accordingly, Fig. 6 provides the same examples used in Figs. 4 and 5 after embedding them into the PIP–style background phase. 4.2. Security Considerations In the proposed scheme, two key considerations must be addressed: resistance to current attacks and difficulty for color-blind bots. These aspects are discussed in the following subsections. 4.2.1. Resistance to current attacks The proposed CAPTCHA system has been designed with an emphasis on resisting prevalent automated attacks, such as adversarial machine learning techniques [ 2 ]. By leveraging the visual complexity inherent in PIP backgrounds—combined with randomized text styles, positions, and color blending—the system introduces a high degree of visual noise and distortion that hinders the effectiveness of these automated solvers. Furthermore, the integration of printed and handwritten text in varying visual configurations significantly increases the challenge for automated tools to accurately segment and interpret characters, thereby enhancing the CAPTCHA's robustness against sophisticated AI-driven attacks. 4.2.2. Difficulty for color-blind bots The use of PIP principles in the CAPTCHA generation introduces an additional security layer specifically designed to exploit limitations in machine vision models that mimic human visual perception. In particular, PIPs are traditionally effective at distinguishing between individuals with normal color vision and those with color vision deficiencies. In this context, their application serves as a means to thwart bots that lack the nuanced color discrimination capabilities of human users. By embedding text in a way that is perceptible only through specific chromatic contrasts, the CAPTCHA ensures that bots—which typically do not emulate color perception subtleties—struggle to isolate and recognize the embedded text, further fortifying the system’s security. 4.3. Usability Considerations The usability aspect of our proposed scheme should be evaluated with a focus on two main factors: human readability and accessibility for individuals with color vision variations. These aspects are explained in the following subsections. 4.3.1. Human readability (tested on different user groups) To ensure the usability of the proposed CAPTCHA system, extensive testing was conducted across diverse user groups, including individuals of varying ages, educational backgrounds, and levels of digital literacy. This will be detailed in the following. The design prioritizes clear text visibility for human users by optimizing font styles, sizes, and color contrasts within the constraints of the PIP format. Despite the deliberate introduction of visual complexity to deter automated recognition, careful balancing was applied to maintain human readability. Feedback from pilot testing guided iterative refinements to improve clarity without compromising security, demonstrating that the CAPTCHA remains usable and user-friendly across a broad demographic. 4.3.2. Color vision variations and accessibility The system also accounts for variations in human color perception, particularly those related to common forms of color vision deficiency. While PIP principles inherently exploit color differentiation to enhance security, design parameters were adjusted to preserve a baseline level of accessibility. Specific color combinations known to be problematic for users with red-green or blue-yellow deficiencies were avoided, in alignment with empirical findings from color vision research. Additionally, the CAPTCHA's difficulty was calibrated to ensure that users with mild to moderate color vision impairments could still successfully complete the challenge, promoting inclusivity without undermining the CAPTCHA’s resistance to bots. 5. Experimental evaluation This section presents an evaluation of the proposed approach from both security and usability perspectives, based on a structured experimental study. The experimental setup and procedures for each aspect are described in detail. 5.1. Experiment setup To assess the usability of the proposed Arabic CAPTCHA generation method, a within-subject laboratory study was conducted. The experiment was carefully designed to evaluate user interaction, readability, and accessibility across different CAPTCHA formats. Participant recruitment began on 15 April 2025 and ended on 30 May 2025. The following subsections detail the key components of the experimental setup. It is important noting that all methods were carried out in accordance with relevant guidelines and regulations. Also, all experimental protocols were approved by a named institutional. Besides, informed consent was obtained from all subjects and/or their legal guardian. 5.1.1. Design A within-subject design was employed to ensure that each participant experienced all CAPTCHA variants under controlled conditions, allowing for direct comparisons of usability factors. 5.1.2. Description of dataset The dataset comprised both meaningful and meaningless Arabic text samples in printed and handwritten forms. Printed samples were generated using the method described in [ 21 ], while meaningless handwritten samples were produced using the tool outlined in [ 38 ]. The generation of meaningful handwritten words followed the approach introduced in [ 3 ]. In total, 60 samples were generated to support the evaluation process. 5.1.3. PIP design parameters: The design of the pseudoisochromatic CAPTCHA plates was carefully configured to balance security and human readability. Several visual parameters were adjusted, each playing a critical role in shaping the user experience and resistance to automated attacks. These parameters are outlined in the following points. Color Schemes A range of foreground and background color combinations was employed to ensure sufficient contrast for users with normal vision while remaining challenging for automated solvers. Color pairings known to be problematic for readability (e.g., white-on-yellow or yellow-on-white, as identified in [ 39 ]) were explicitly avoided. Instead, empirically validated combinations were used to maintain visual distinction and legibility. Noise Density Background noise elements—random circles varying in chromaticity, luminance, and size—were introduced to obscure the CAPTCHA text from machine-based recognition systems. The noise level was carefully calibrated to avoid excessive interference with human readability, particularly for users with mild visual impairments. Distractor Types Additional distractor elements, including overlapping shapes and color gradients, were incorporated to further complicate segmentation and character recognition by bots. Figure 7 illustrates a sample of implementing this parameter. These distractors were randomized per image to reduce pattern predictability and increase variability across samples. Collectively, these parameters were optimized to create visually complex CAPTCHA images that remain accessible to human users while significantly hindering machine recognition capabilities. 5.1.4. Participant demographics To capture relevant background information, participants were asked to complete a brief demographic survey prior to beginning the experiment. The survey included the following items: prior experience with Arabic CAPTCHAs (i.e., whether the participant had previously encountered or used Arabic CAPTCHAs), gender, age group, and education level. This information was collected to analyze potential variations in usability performance and perception based on user characteristics, and to ensure a diverse and representative sample of Arabic-speaking users. 5.1.5. Baseline CAPTCHA methods for comparison To evaluate the impact of the PIP embedding, the original generated Arabic text samples—both printed and handwritten—were used as baseline CAPTCHA methods, as shown previously in Figs. 3 and 4. These unembedded samples represent standard, high-contrast text images without any background noise or color blending. By comparing user responses and bot resistance between these plain text images and their PIP-enhanced counterparts, we were able to quantify the trade-offs in usability and security introduced by our proposed method. Specifically, the main reason behind using the original generated samples is as follows. Control Variable The original samples (plain printed or handwritten text) serve as a controlled version of the CAPTCHA, isolating the effect of the PIP-based embedding. Usability Benchmark Comparing user performance (e.g., recognition accuracy and time) on plain samples versus PIP-enhanced samples helps quantify any usability trade-offs. Security Contrast Bots may find the original text easier to segment and recognize, so the baseline also allows a direct assessment of how much the PIP-based method enhances resistance to automated attacks. 5.1.6. System An online testing interface was developed using Weavely Forms and the Figma platform, providing participants with interactive access to the proposed CAPTCHA samples and recording their performance metrics and feedback. 5.2. Experiment procedure This section outlines the detailed methodology of the experiment, including participant instructions, procedures for evaluating both security and usability, and the data collected. i. Instructions to participants : At the outset, participants completed a demographic survey capturing relevant background information. They were then informed that they would be presented with 60 Arabic text-based CAPTCHA samples displayed in the following order: (1) printed-based meaningful and meaningless texts, then (2) handwritten-based meaningful and meaningless texts, alongside baseline samples representing each one of them. Upon navigating to the CAPTCHA challenge page, participants were instructed to carefully observe each displayed CAPTCHA and enter the recognized characters into the provided input field, submitting their response by clicking the submit button. It is important to note that no feedback regarding answer correctness or hints was provided during the task. Participants were also advised to minimize distractions and maintain focus throughout the experiment. ii. Security experiment procedure To assess the security robustness of the proposed CAPTCHA system, we utilized the Google Vision API [ 40 ] as a benchmark for evaluating resistance against recognition and segmentation attacks. Given its widespread use and high accuracy, the Cloud Vision API served as an appropriate tool for this purpose. All 60 CAPTCHA samples were processed through the Google Vision engine, and the resulting text labels were categorized into four recognition outcomes: completely recognized, partially recognized, incorrectly recognized, and not recognized. Table 2 summarizes these categories along with their descriptions. Table 2 Recognition Outputs Categories Recognition Category (RC) Description Completely (C) All characters correctly recognized Partially (P) Some characters correctly recognized Incorrectly (I) All characters incorrectly recognized Not recognized (NR) No characters recognized iii. Usability experiment procedure : The usability of the proposed CAPTCHA system was evaluated through both quantitative and qualitative metrics. Quantitative evaluation involved measuring the efficiency and effectiveness of the system. Efficiency was quantified by recording the time (in seconds) taken by participants to accurately type the CAPTCHA text—from clicking the ‘Start’ button to submitting their response. Effectiveness was measured as the success rate, defined by the proportion of CAPTCHAs correctly solved by the participants. For the qualitative metric, data were obtained via a post-experiment survey assessing user satisfaction and perceived difficulty. iv. Collected data : During the experiment, key performance indicators including success rates, response times, and user satisfaction scores were systematically recorded and analyzed to provide a comprehensive assessment of the proposed approach. 6. Results In the experimental study, all participants completed the tasks successfully. The security evaluation, which shows the robustness level against such attacks, of proposed approach is presented. This is followed by the usability results which includes the efficiency, effectiveness, and satisfaction rates of the participants. 6.1. Security evaluation results Table 3 summarizes the total percentage of robustness level of the proposed approach against such sophisticated attacks using Google Vision API [ 40 ]. In particular, the results showed the robustness level of printed-based for meaningful, meaningless and baseline text. Also the robustness level of handwritten-based for meaningful, meaningless and baseline texts are shown. Table 3 The results of security evaluation RC Printed-based text Handwritten-based text Meaningful Meaningless Baseline Meaningful Meaningless Baseline Meaningful Meaningless Meaningful Meaningless C 0% 0% 90% 100% 0 0 70% 60% P 0% 0% 10% 0% 0 0 20% 40% I 0% 0% 0% 0% 0 0 0% 0% N 100% 100% 0% 0% 100% 100% 10% 0% 6.2. Usability evaluation results The usability experiment is performed after the security experiment. This section shows the results of testing the efficiency, effectiveness, and the satisfaction with the proposed approach. 6.2.1. Efficiency : Fig. 8 illustrates the average duration (in seconds) participants required to correctly input the CAPTCHA text. The bars represent the average time taken (in seconds) to solve each CAPTCHA type. The dashed lines show the overall average efficiency: 7.75 seconds for the Baseline scheme and 23.75 seconds for the Proposed scheme. The increased time reflects the enhanced robustness and complexity introduced in the Proposed CAPTCHA design. 6.2.2 Effectiveness: The success rate of correctly solving of released CAPTCHA samples is shown in Fig. 9 . The vertical bars represent the average success rates for each condition, while the dashed lines indicate the overall average effectiveness for the Baseline (99.0%) and Proposed (90.8%) schemes. The results highlight a slight decrease in effectiveness for the Proposed scheme due to increased robustness, yet maintain high human solvability across all text types. 6.2.3. Satisfaction : Upon completing all assigned tasks, participants were asked to respond to a set of questions assessing their experience with the proposed CAPTCHA approach. The survey questions were grouped into four key categories: Perceived Readability and Clarity , Ease of Solving and Interaction , User Frustration or Fatigue , and Aesthetic and Visual Impression . In the first category, participants were asked questions such as: "It was easy to read the characters in the CAPTCHA images" and "The contrast between the text and background was sufficient for me to read the CAPTCHA." For Ease of Solving and Interaction , the question was: "I was able to recognize and type the CAPTCHA text without much effort." The third category, User Frustration or Fatigue , included: "Solving the CAPTCHAs made me feel frustrated or fatigued." Finally, under Aesthetic and Visual Impression , the statement was: "The design of the CAPTCHA was visually appealing." Responses were rated on a five-point Likert scale ranging from 1 ( Strongly Disagree ) to 5 ( Strongly Agree ), represented by a star rating system. A summary of the survey results is presented in Fig. 10. 7. Discussion This section synthesizes the findings of the experimental evaluation and reflects on the broader implications of the proposed PIP-style Arabic CAPTCHA approach. By analyzing both security and usability outcomes, we aim to assess the effectiveness of the method and identify its practical strengths and limitations. Furthermore, we discuss specific considerations relevant to Arabic-language contexts and propose directions for future enhancement. The following subsections explore key trade-offs, advantages, constraints, and opportunities for improvement. 7.1. Trade-offs between security and usability The findings of this study highlight an inherent trade-off between CAPTCHA security and usability. The integration of PIP-style backgrounds significantly enhances resistance to automated recognition systems, particularly by impeding character segmentation and distorting character shapes. However, this enhancement comes at the cost of reduced usability for human users. For instance, handwritten text samples embedded in PIP backgrounds demonstrated robust security performance but required longer response times and exhibited slightly lower success rates among participants, as illustrated in Fig. 8 . In contrast, both meaningful and meaningless printed texts offered a more user-friendly experience while still maintaining substantial resistance to segmentation and recognition attacks. Given that printed text samples outperformed handwritten samples in terms of both efficiency and effectiveness, their use in the proposed scheme is recommended. Nonetheless, further research is needed to validate these findings and explore potential optimizations. Overall, achieving an optimal balance between security and usability remains a central challenge in CAPTCHA design, particularly for languages with complex scripts such as Arabic. 7.2. Strengths of the PIP approach in Arabic contexts The proposed approach capitalizes on the visual and linguistic intricacies of the Arabic script, which already poses challenges to OCR systems. When coupled with the PIP framework, these inherent complexities are further amplified, leading to stronger protection against bot-based attacks, as illustrated in Table 3 . Moreover, the stylistic variations introduced through randomized fonts, character sizes, and chromatic noise patterns enhance both unpredictability and visual entropy. In user testing, Arabic-speaking participants expressed a positive response toward the novel presentation style, noting its cultural and linguistic relevance, as summarized in Fig. 10. This underscores the adaptability of the PIP mechanism to Arabic-language applications, where traditional Latin-based CAPTCHA systems often fall short in both usability and effectiveness. 7.3. Limitations: potential issues for color-blind users, complex generation process Despite its strengths, the PIP-based CAPTCHA system is not without limitations. One critical concern is accessibility for users with color vision deficiencies. Although the design avoids color combinations known to be problematic, the inherently chromatic nature of PIP backgrounds can still impede recognition for color-blind users. Future iterations should include an accessibility mode or offer contrast-enhancing alternatives. Additionally, the generation process is relatively complex, involving fine-tuned color selection, shape rendering, and integration with character layers. This complexity may pose implementation challenges for web developers or system integrators lacking advanced image processing capabilities. 7.4. Future enhancements To address these limitations and extend the system’s applicability, several enhancements are proposed. First, the development of dynamic plate generators —capable of adjusting color schemes and visual density based on real-time user performance—could improve accessibility and personalization. Second, adaptive contrast techniques could be integrated to detect and compensate for potential vision deficiencies, ensuring that CAPTCHA elements remain distinguishable across a wider user base. Lastly, further research into multimodal CAPTCHA systems , combining PIP visuals with auditory or haptic cues, may provide inclusive alternatives while preserving high security standards. 8. Conclusion This study introduced a novel Arabic CAPTCHA generation method that incorporates Pseudoisochromatic Plate (PIP) principles to improve resistance against automated recognition attacks while maintaining human readability. The proposed framework demonstrated strong robustness, particularly through the integration of visual complexity via color blending and randomized character representation. Experimental results confirmed the system’s effectiveness in countering machine-based attacks, including those leveraging advanced recognition tools like the Google Vision API. Furthermore, usability evaluations revealed that while the handwritten PIP-embedded CAPTCHAs offer heightened security, printed text variants provide superior usability in terms of efficiency and effectiveness. This trade-off emphasizes the ongoing challenge of balancing security and accessibility in CAPTCHA design—especially for complex scripts such as Arabic. The proposed system offers a significant contribution to the field by introducing perceptual-based obfuscation mechanisms that are linguistically and culturally adapted. Future research will focus on enhancing accessibility for users with color vision deficiencies and refining dynamic generation techniques to support broader inclusivity and implementation in real-world applications. Declarations Funding: The Researcher would like to thank the Deanship of Graduate Studies and Scientific Research at Qassim University for financial support (QU-APC-2025). Author Contribution Conceptualization, S. A. Alsuhibany.; methodology, S. A. Alsuhibany.; software, S. A. Alsuhibany.; validation, S. A. Alsuhibany.; formal analysis, S. A. Alsuhibany.; investigation, S. A. Alsuhibany.; resources, S. A. Alsuhibany.; data curation, S. A. Alsuhibany.; writing—original draft preparation, S. A. Alsuhibany.; writing—review and editing, S. A. Alsuhibany.; visualization, S. A. Alsuhibany Acknowledgement The Researcher would like to thank the Deanship of Graduate Studies and Scientific Research at Qassim University for financial support (QU-APC-2025) Data Availability The data that support the findings of this study are available from the corresponding author, Suliman A. Alsuhibany, upon reasonable request. References Alsuhibany, S. A. & Parvez, M. T. Secure Arabic Handwritten CAPTCHA Generation Using OCR Operations. In: 2016 15th International Conference on Frontiers in Handwriting Recognition (ICFHR), pp. 126–131. (2016). https://doi.org/10.1109/ICFHR.2016.31 Alsuhibany, S. A. A survey on adversarial perturbations and attacks on CAPTCHAs. Appl. Sci. 13 (7), 4602. https://doi.org/10.3390/app13074602 (2023). Alrasheed, G. & Alsuhibany, S. A. Enhancing Security of Online Interfaces: Adversarial Handwritten Arabic CAPTCHA Generation. Appl. Sci. 15 (6), 2972. https://doi.org/10.3390/app15062972 (2025). Shirali-Shahreza, M. H. & Shirali-Shahreza, M. Persian/Arabic Baffletext CAPTCHA J. UCS 12 (12), 1783–1796. (2006). Shahreza, M. S. Verifying Spam SMS by Arabic CAPTCHA. In: 2nd Int. Conf. on Information & Communication Technologies (ICT), vol. 1, pp. 78–83. (2006). https://doi.org/10.1109/ICT.2006.253141 Khan, B., Alghathbar, K. S., Khan, M. K., AlKelabi, A. M. & AlAjaji, A. Using Arabic CAPTCHA for cyber security. In: Security Technology, Disaster Recovery and Business Continuity, pp. 8–17. Springer. (2010). https://doi.org/10.1007/978-3-642-16644-0_2 Bakry, M., Khamis, M., Abdennadher, S. & Alhodathi, H. Arecaptcha: Outsourcing Arabic text digitization to native speakers. In: 11th IAPR Int. Workshop on Document Analysis Systems (DAS), pp. 304–308. (2014). https://doi.org/10.1109/DAS.2014.66 Abubaker, H., Salah, K., Al-Muhairi, H. & Bentiba, A. Cloud-based Arabic reCAPTCHA service: Design and architecture. In: AICCSA, pp. 1–6. (2015). https://doi.org/10.1109/AICCSA.2015.7507200 Abubaker, H., Salah, K., Al-Muhairi, H. & Bentiba, A. Architectural design of a cloud-based reCAPTCHA service. In: 12th Int. Conf. on Innovations in Information Technology (IIT), pp. 1–6. (2016). https://doi.org/10.1109/INNOVATIONS.2016.7880041 Rodriguez-Carmona, M. Pseudoisochromatic Plates. In: (ed Luo, R.) Encyclopedia of Color Science and Technology, Springer, Berlin, Heidelberg, 1–2. (2014). Parvez, M. T. & Alsuhibany, S. A. Challenges and opportunities for Arabic CAPTCHAs. Multimed Tools Appl. 83 (5), 14047–14062. https://doi.org/10.1007/s11042-023-15847-9 (2024). Alsuhibany, S. A. & Parvez, M. T. Attack-filtered interactive Arabic CAPTCHAs. J. Inf. Secur. Appl. 70 , 103318. https://doi.org/10.1016/j.jisa.2022.103318 (2022). Khan, B., Alghathbar, K., Khan, M. K., AlKelabi, A. M. & Alajaji, A. Cyber security using Arabic CAPTCHA scheme. Int. Arab. J. Inf. Technol. 10 (1), 76–84 (2013). Banday, M. T. & Sheikh, S. A. CAPTCHA Chall. Int. J. Web Portals 7 (1), 1–27. (2015). Sulaiman, M. T. & Hassan, N. F. Generate Random Arabic Characters and Numbers for CAPTCHA. Iraqi J. Sci. 58 (4B), 2427–2437 (2017). Alsuhibany, S. A. & Alhodathi, H. A. A Parametric Study of Arabic Text-Based CAPTCHA Difficulty for Humans. Intell. Autom. Soft Comput. 31 (1), 523–537. https://doi.org/10.32604/iasc.2022.017266 (2022). Sulaiman, M. T. Propose an Arabic CAPTCHA System based on Chaotic Maps. Eng. Technol. J. 36 (1B), 13–21 (2018). Alsuhibany, S. A., Parvez, M. T., Alrobah, N., Almohaimeed, F. & Alduayji, S. Evaluating robustness of Arabic CAPTCHAs. In: 2nd Int. Conf. on Anti-Cyber Crimes (ICACC), pp. 81–86. (2017). https://doi.org/10.1109/ICACC.2017.7914984 Shirali-Shahreza, M. H. & Shirali-Shahreza, M. Multilingual CAPTCHA. In: 2007 IEEE Int. Conf. on Computational Cybernetics, pp. 135–139. (2007). Shirali-Shahreza, M. H. & Shirali-Shahreza, M. Advanced Nastaliq CAPTCHA. In: 7th IEEE Int. Conf. on Cybernetic Intelligent Systems, pp. 1–3. (2008). Alsuhibany, S. A. & Alquraishi, M. Usability and Security of Arabic Text-based CAPTCHA Using Visual Cryptography. Comput. Syst. Sci. Eng. 40 (2), 421–440 (2022). Parvez, M. T., Alsuhibani, A. M. & Alamri, A. H. Educational and Cybersecurity Applications of an Arabic CAPTCHA Gamification System. Ingénierie des Systèmes. d’Information 28 (5), 1275 (2023). Lajmi, H., Idoudi, F., Njah, H., Kammoun, H. M. & Njah, I. Strengthening Applications' Security with Handwritten Arabic Calligraphy CAPTCHA. In: 8th IEEE Forum on Research and Technologies for Society and Industry Innovation (RTSI), pp. 548–553. (2024). Aldosari, M. H. Innovative multilingual CAPTCHA Based on handwritten characteristics. In: Multilingualism and Bilingualism. IntechOpen. (2017). Parvez, M. T. & Alsuhibany, S. A. Segmentation-Validation based Handwritten Arabic CAPTCHA Generation. Comput. Secur. 96 , 101829. https://doi.org/10.1016/j.cose.2020.101829 (2020). Aldosari, M. H. & Al-Daraiseh, A. A. Strong multilingual CAPTCHA based on handwritten characters. In: 7th Int. Conf. on Information and Communication Systems (ICICS), pp. 239–245. (2016). Fawa’reh, M., Qasaimeh, M., AbuArja, I. & Al-Fayoumi, M. Mitigating Deep learning Attacks Against Text Image CAPTCHA Using Arabic Scheme. Int. J. Commun. Antenna Propag. 11 (4), 288–296 (2021). Alsuhibany, S. A. & Alnoshan, A. A. Interactive Handwritten and Text-Based Handwritten Arabic CAPTCHA Schemes for Mobile Devices: A Comparative Study. IEEE Access. 9 , 140991–141001 (2021). Abubaker, H., Salah, K., Al-Muhairi, H. & Bentiba, A. Digital Arabic content: challenges and opportunities. In: Int. Conf. on ICT Research (ICTRC), pp. 330–333. (2015). Abubaker, H., Salah, K., Al-Muhairi, H. & Bentiba, A. Arabic reCAPTCHA Service for Enhancing Digitization of Arabic Manuscripts. Arab. J. Sci. Eng. 42 (8), 3391–3408 (2017). Abubaker, H. B. Arabic reCAPTCHA Service for Web Security and Old Arabic Manuscripts Digitization. Doctoral Dissertation, Khalifa Univ. Sci. Technol. Res. (2016). Birch, J. Diagnosis of Defective Colour Vision second edn (Oxford University Press, 2012). Cole, B. L. Assessment of inherited colour vision defects in clinical practice. Clin. Exp. Optom. 87 (4–5), 276–293 (2004). Hurvich, L. M. & Jameson, D. An opponent-process theory of color vision. Psychol. Rev. 64 (6), 384–404 (1957). Ishihara, S. Tests for Colour Blindness. Kanehara Shuppan (1972). Hardy, G. H., Rand, G. & Rittler, M. C. HRR Pseudoisochromatic Plates. Richmond Products. (1954). Mollon, J. D. & Reffin, J. P. The Cambridge Colour Test. J. Opt. Soc. Am. A . 6 (3), 586–597 (1989). Alsuhibany, S. A., Almohaimeed, F. N. & Alrobah, N. A. Synthetic Arabic handwritten CAPTCHA. Int. J. Inf. Comput. Secur. 16 (3–4), 385–398 (2021). Bursztein, E. et al. Easy does it: More usable CAPTCHAs. In: Proc. SIGCHI Conf. Hum. Factors Comput. Syst., pp. 2637–2646. (2014). Google Cloud Cloud Vision API. (2022). https://cloud.google.com/vision (accessed 10 June 2025). Additional Declarations No competing interests reported. Cite Share Download PDF Status: Posted Version 1 posted You are reading this latest preprint version Research Square lets you share your work early, gain feedback from the community, and start making changes to your manuscript prior to peer review in a journal. As a division of Research Square Company, we’re committed to making research communication faster, fairer, and more useful. We do this by developing innovative software and high quality services for the global research community. Our growing team is made up of researchers and industry professionals working together to solve the most critical problems facing scientific publishing. Also discoverable on Platform About Our Team In Review Editorial Policies Advisory Board Help Center Resources Author Services Accessibility API Access RSS feed Manage Cookie Preferences © Research Square 2026 | ISSN 2693-5015 (online) Privacy Policy Terms of Service Do Not Sell My Personal Information {"props":{"pageProps":{"initialData":{"identity":"rs-7367912","acceptedTermsAndConditions":true,"allowDirectSubmit":true,"archivedVersions":[],"articleType":"Article","associatedPublications":[],"authors":[{"id":517975604,"identity":"16d3c975-7ba6-4036-b4ab-29ee74846dcc","order_by":0,"name":"Suliman A. Alsuhibany","email":"data:image/png;base64,iVBORw0KGgoAAAANSUhEUgAAAZAAAAAyAQMAAABI0h/eAAAABlBMVEX///8AAABVwtN+AAAACXBIWXMAAA7EAAAOxAGVKw4bAAAA90lEQVRIiWNgGAWjYFAC5gYGxgaJBCCL8QFY4ABBLYxwLcwGpGhhAGlhkyBKi3z7webPvDss8vhn95hVvG1jkOO7kcD84QceLQZnEtukec9IFEvcOWN2c24bg7HkjQQ2yR58WhgS25h52yQSG27kmN3mbWNI3ADUwsCDz2H9D4EOA2qZD9RSDNRSD9TC/PEPPs/cSGyQBmnZANQCtI4hweBGAoM0PlsMbjxsk5zbJlFseCOtWHLOOQnDmWcetknL4HVY8uEPb9vq8uRuJG/88KbMRp7vePLhj2/wOQwBOAyAvgZFDTCaiATsD/AG1CgYBaNgFIxcAAAxClGkX+upnAAAAABJRU5ErkJggg==","orcid":"","institution":"Qassim University","correspondingAuthor":true,"prefix":"","firstName":"Suliman","middleName":"A.","lastName":"Alsuhibany","suffix":""}],"badges":[],"createdAt":"2025-08-13 20:08:22","currentVersionCode":1,"declarations":"","doi":"10.21203/rs.3.rs-7367912/v1","doiUrl":"https://doi.org/10.21203/rs.3.rs-7367912/v1","draftVersion":[],"editorialEvents":[],"editorialNote":"","failedWorkflow":false,"files":[{"id":92023505,"identity":"1fb725b7-f463-408d-b222-36a5342df259","added_by":"auto","created_at":"2025-09-23 18:19:13","extension":"docx","order_by":0,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":3650514,"visible":true,"origin":"","legend":"","description":"","filename":"Paperfinal.docx","url":"https://assets-eu.researchsquare.com/files/rs-7367912/v1/3808ed58bff1a9f33430874f.docx"},{"id":92023496,"identity":"7b90a4dd-9d05-44ff-a35f-b952050be03e","added_by":"auto","created_at":"2025-09-23 18:19:13","extension":"json","order_by":1,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":3785,"visible":true,"origin":"","legend":"","description":"","filename":"2156da1fbb8e45b29ea9f785f69b721e.json","url":"https://assets-eu.researchsquare.com/files/rs-7367912/v1/4b3917aebbad65cb5f354414.json"},{"id":92024146,"identity":"1adc4b58-84b4-4784-8fb1-5b8c4ebb3187","added_by":"auto","created_at":"2025-09-23 18:35:13","extension":"xml","order_by":2,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":121861,"visible":true,"origin":"","legend":"","description":"","filename":"2156da1fbb8e45b29ea9f785f69b721e1enriched.xml","url":"https://assets-eu.researchsquare.com/files/rs-7367912/v1/c3a1f0357795af85d041348c.xml"},{"id":92023126,"identity":"a0b40cfa-b8b8-4080-b3a7-2ebc25863545","added_by":"auto","created_at":"2025-09-23 18:11:13","extension":"png","order_by":3,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":31255,"visible":true,"origin":"","legend":"","description":"","filename":"floatimage1.png","url":"https://assets-eu.researchsquare.com/files/rs-7367912/v1/4e7232904eff61f408b1b999.png"},{"id":92023503,"identity":"0ad09c47-6f7c-4f18-b315-139950548a1b","added_by":"auto","created_at":"2025-09-23 18:19:13","extension":"png","order_by":4,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":3416,"visible":true,"origin":"","legend":"","description":"","filename":"floatimage10.png","url":"https://assets-eu.researchsquare.com/files/rs-7367912/v1/3d7641f483740d00216e1a77.png"},{"id":92023497,"identity":"5a92a195-a84c-4778-ba0f-c178b512bf34","added_by":"auto","created_at":"2025-09-23 18:19:13","extension":"png","order_by":5,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":6186,"visible":true,"origin":"","legend":"","description":"","filename":"floatimage11.png","url":"https://assets-eu.researchsquare.com/files/rs-7367912/v1/289f8d0488b3fde5e1e243ab.png"},{"id":92023141,"identity":"b5b637e6-ba68-42d1-ae46-9cf5179f7c15","added_by":"auto","created_at":"2025-09-23 18:11:13","extension":"png","order_by":6,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":96382,"visible":true,"origin":"","legend":"","description":"","filename":"floatimage12.png","url":"https://assets-eu.researchsquare.com/files/rs-7367912/v1/707e5a74ca04c0bf2f0d1c3c.png"},{"id":92023136,"identity":"6be2fc56-cd7a-4bba-80b6-ae7aad7f172d","added_by":"auto","created_at":"2025-09-23 18:11:13","extension":"png","order_by":7,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":93190,"visible":true,"origin":"","legend":"","description":"","filename":"floatimage13.png","url":"https://assets-eu.researchsquare.com/files/rs-7367912/v1/35a75f39b3906ba28de34f89.png"},{"id":92023138,"identity":"e0cd279a-64d3-49df-bba9-0c58c821f74f","added_by":"auto","created_at":"2025-09-23 18:11:13","extension":"png","order_by":8,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":100139,"visible":true,"origin":"","legend":"","description":"","filename":"floatimage14.png","url":"https://assets-eu.researchsquare.com/files/rs-7367912/v1/1e9297ac527963e574a2a9d4.png"},{"id":92023134,"identity":"0e105e3f-96de-47a8-ac0a-4ad574d59c6b","added_by":"auto","created_at":"2025-09-23 18:11:13","extension":"png","order_by":9,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":107315,"visible":true,"origin":"","legend":"","description":"","filename":"floatimage15.png","url":"https://assets-eu.researchsquare.com/files/rs-7367912/v1/b8b262952402270bbc8a76fd.png"},{"id":92023512,"identity":"ad7eef04-2d2b-4b98-8bdd-cbe01ddcfb9f","added_by":"auto","created_at":"2025-09-23 18:19:14","extension":"png","order_by":10,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":66790,"visible":true,"origin":"","legend":"","description":"","filename":"floatimage16.png","url":"https://assets-eu.researchsquare.com/files/rs-7367912/v1/be412db9274ab9e03df12894.png"},{"id":92023155,"identity":"6e9375ec-d063-4847-863b-fcd8fde25b10","added_by":"auto","created_at":"2025-09-23 18:11:14","extension":"png","order_by":11,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":84090,"visible":true,"origin":"","legend":"","description":"","filename":"floatimage17.png","url":"https://assets-eu.researchsquare.com/files/rs-7367912/v1/9a3131f698b822149492164e.png"},{"id":92023504,"identity":"41cf027a-53b7-423f-a630-4addceea02cb","added_by":"auto","created_at":"2025-09-23 18:19:13","extension":"png","order_by":12,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":76117,"visible":true,"origin":"","legend":"","description":"","filename":"floatimage18.png","url":"https://assets-eu.researchsquare.com/files/rs-7367912/v1/da38006e006486381bfe701c.png"},{"id":92023143,"identity":"261f991e-6e2e-4c57-a977-df5a542ae4cb","added_by":"auto","created_at":"2025-09-23 18:11:13","extension":"png","order_by":13,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":6597,"visible":true,"origin":"","legend":"","description":"","filename":"floatimage19.png","url":"https://assets-eu.researchsquare.com/files/rs-7367912/v1/5d33da0592cb3203eb66cc17.png"},{"id":92023150,"identity":"fbe03399-e73d-4b7e-b1df-c72b22d92eaf","added_by":"auto","created_at":"2025-09-23 18:11:14","extension":"png","order_by":14,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":30173,"visible":true,"origin":"","legend":"","description":"","filename":"floatimage2.png","url":"https://assets-eu.researchsquare.com/files/rs-7367912/v1/4f5e4be0244d8c96dfc3f6f8.png"},{"id":92024147,"identity":"0dbb95b0-4d65-4572-ad65-3dc361979bc3","added_by":"auto","created_at":"2025-09-23 18:35:14","extension":"png","order_by":15,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":6315,"visible":true,"origin":"","legend":"","description":"","filename":"floatimage20.png","url":"https://assets-eu.researchsquare.com/files/rs-7367912/v1/642dd9b44f11550fc63291b2.png"},{"id":92023144,"identity":"d9b97459-0a91-4994-9c7e-bb68ed184acc","added_by":"auto","created_at":"2025-09-23 18:11:13","extension":"png","order_by":16,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":5953,"visible":true,"origin":"","legend":"","description":"","filename":"floatimage21.png","url":"https://assets-eu.researchsquare.com/files/rs-7367912/v1/433230bedf43c0550227330f.png"},{"id":92023515,"identity":"4570e5e8-779f-43dd-a29e-0ca94905bc0f","added_by":"auto","created_at":"2025-09-23 18:19:14","extension":"png","order_by":17,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":6436,"visible":true,"origin":"","legend":"","description":"","filename":"floatimage22.png","url":"https://assets-eu.researchsquare.com/files/rs-7367912/v1/c8d2342c371ec91257abce3d.png"},{"id":92023507,"identity":"ff20e8e8-16ff-4b9a-b3df-beb3de9a0a5c","added_by":"auto","created_at":"2025-09-23 18:19:14","extension":"png","order_by":18,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":6269,"visible":true,"origin":"","legend":"","description":"","filename":"floatimage23.png","url":"https://assets-eu.researchsquare.com/files/rs-7367912/v1/f3a616126c2102c611f733d6.png"},{"id":92023979,"identity":"710114ef-6cd5-40c8-a225-2be6e37d6990","added_by":"auto","created_at":"2025-09-23 18:27:14","extension":"jpeg","order_by":19,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":643890,"visible":true,"origin":"","legend":"","description":"","filename":"floatimage3.jpeg","url":"https://assets-eu.researchsquare.com/files/rs-7367912/v1/118d9d504adf46a4c90f0b96.jpeg"},{"id":92023977,"identity":"5478d0e7-1548-4e24-b012-f6635e0d728c","added_by":"auto","created_at":"2025-09-23 18:27:14","extension":"gif","order_by":20,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":30970,"visible":true,"origin":"","legend":"","description":"","filename":"floatimage4.gif","url":"https://assets-eu.researchsquare.com/files/rs-7367912/v1/3cc42c49ac412aef68baaee1.gif"},{"id":92023179,"identity":"04b69776-f6cb-448d-ab00-6aec9612c9c6","added_by":"auto","created_at":"2025-09-23 18:11:14","extension":"png","order_by":21,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":1028665,"visible":true,"origin":"","legend":"","description":"","filename":"floatimage5.png","url":"https://assets-eu.researchsquare.com/files/rs-7367912/v1/34e0797c36f574fab3353f1c.png"},{"id":92023153,"identity":"50403e98-7fd6-4b23-9349-2f8aee92d53d","added_by":"auto","created_at":"2025-09-23 18:11:14","extension":"jpeg","order_by":22,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":370874,"visible":true,"origin":"","legend":"","description":"","filename":"floatimage6.jpeg","url":"https://assets-eu.researchsquare.com/files/rs-7367912/v1/a7ebf3d98df46d50363447c6.jpeg"},{"id":92023160,"identity":"f1437612-03e9-42ca-b0f0-fc136904f96d","added_by":"auto","created_at":"2025-09-23 18:11:14","extension":"png","order_by":23,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":541118,"visible":true,"origin":"","legend":"","description":"","filename":"floatimage7.png","url":"https://assets-eu.researchsquare.com/files/rs-7367912/v1/69458a60d6a4640c8cb05327.png"},{"id":92023509,"identity":"d47da302-040d-43d9-b9a2-ad6a07aff09d","added_by":"auto","created_at":"2025-09-23 18:19:14","extension":"png","order_by":24,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":3340,"visible":true,"origin":"","legend":"","description":"","filename":"floatimage8.png","url":"https://assets-eu.researchsquare.com/files/rs-7367912/v1/4fcf51623066bcb46cb43c1a.png"},{"id":92023506,"identity":"d787ed19-2215-4494-9d13-dee1892bf5a6","added_by":"auto","created_at":"2025-09-23 18:19:14","extension":"png","order_by":25,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":3031,"visible":true,"origin":"","legend":"","description":"","filename":"floatimage9.png","url":"https://assets-eu.researchsquare.com/files/rs-7367912/v1/ca8cb8eb2e7c575394203bd1.png"},{"id":92023166,"identity":"27839a39-f872-41f0-ba83-02d8efd072e1","added_by":"auto","created_at":"2025-09-23 18:11:14","extension":"png","order_by":26,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":11489,"visible":true,"origin":"","legend":"","description":"","filename":"Onlinefloatimage1.png","url":"https://assets-eu.researchsquare.com/files/rs-7367912/v1/f691a01b4485a5fe5c9fcc47.png"},{"id":92023169,"identity":"e6ea4e43-147d-4e09-acaa-b415e18c6587","added_by":"auto","created_at":"2025-09-23 18:11:14","extension":"png","order_by":27,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":1857,"visible":true,"origin":"","legend":"","description":"","filename":"Onlinefloatimage10.png","url":"https://assets-eu.researchsquare.com/files/rs-7367912/v1/affd94adf8785ea754e18ed8.png"},{"id":92023157,"identity":"7dc5a223-378a-4adc-ad4c-d289248934a5","added_by":"auto","created_at":"2025-09-23 18:11:14","extension":"png","order_by":28,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":4083,"visible":true,"origin":"","legend":"","description":"","filename":"Onlinefloatimage11.png","url":"https://assets-eu.researchsquare.com/files/rs-7367912/v1/ce46d69f447d946dfad05488.png"},{"id":92023518,"identity":"c7be6117-84db-4e72-a5e2-9e31e288f836","added_by":"auto","created_at":"2025-09-23 18:19:14","extension":"png","order_by":29,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":29145,"visible":true,"origin":"","legend":"","description":"","filename":"Onlinefloatimage12.png","url":"https://assets-eu.researchsquare.com/files/rs-7367912/v1/498ac69992cc6957098289be.png"},{"id":92023164,"identity":"25a7f015-4206-4e4a-b4c1-6f4229855fcd","added_by":"auto","created_at":"2025-09-23 18:11:14","extension":"png","order_by":30,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":30891,"visible":true,"origin":"","legend":"","description":"","filename":"Onlinefloatimage13.png","url":"https://assets-eu.researchsquare.com/files/rs-7367912/v1/fd97391c1566eb1124b9cd90.png"},{"id":92023159,"identity":"477cda93-b456-462b-baa1-7286b6df0633","added_by":"auto","created_at":"2025-09-23 18:11:14","extension":"png","order_by":31,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":61868,"visible":true,"origin":"","legend":"","description":"","filename":"Onlinefloatimage14.png","url":"https://assets-eu.researchsquare.com/files/rs-7367912/v1/0de3242a10fadff79bdcdd3f.png"},{"id":92023981,"identity":"bc6e2d82-7c23-4487-98d5-ae5c684558c6","added_by":"auto","created_at":"2025-09-23 18:27:14","extension":"png","order_by":32,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":64120,"visible":true,"origin":"","legend":"","description":"","filename":"Onlinefloatimage15.png","url":"https://assets-eu.researchsquare.com/files/rs-7367912/v1/10927226db86bb619683b6e3.png"},{"id":92023520,"identity":"2dc8dfd7-e508-47d9-8e65-9d384c3ca68c","added_by":"auto","created_at":"2025-09-23 18:19:14","extension":"png","order_by":33,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":13689,"visible":true,"origin":"","legend":"","description":"","filename":"Onlinefloatimage16.png","url":"https://assets-eu.researchsquare.com/files/rs-7367912/v1/bd5c0c84bf327cd7e015827d.png"},{"id":92023165,"identity":"5c7de2cf-92a5-4238-9dc0-779bb6c662a8","added_by":"auto","created_at":"2025-09-23 18:11:14","extension":"png","order_by":34,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":20020,"visible":true,"origin":"","legend":"","description":"","filename":"Onlinefloatimage17.png","url":"https://assets-eu.researchsquare.com/files/rs-7367912/v1/f8124312b37ad10e3881b1d8.png"},{"id":92023980,"identity":"22ddb12e-157c-4f8e-b005-fe6c91da0973","added_by":"auto","created_at":"2025-09-23 18:27:14","extension":"png","order_by":35,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":17981,"visible":true,"origin":"","legend":"","description":"","filename":"Onlinefloatimage18.png","url":"https://assets-eu.researchsquare.com/files/rs-7367912/v1/9f81efb5e3081b691212a589.png"},{"id":92023514,"identity":"4eb424c8-b3be-40bb-b4a2-cd5ddb40ebce","added_by":"auto","created_at":"2025-09-23 18:19:14","extension":"png","order_by":36,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":2174,"visible":true,"origin":"","legend":"","description":"","filename":"Onlinefloatimage19.png","url":"https://assets-eu.researchsquare.com/files/rs-7367912/v1/dfa99d25b49c6fbe7c6473d4.png"},{"id":92023983,"identity":"5bbedd1e-2949-4f4e-9e08-01a0f4e76476","added_by":"auto","created_at":"2025-09-23 18:27:14","extension":"png","order_by":37,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":12236,"visible":true,"origin":"","legend":"","description":"","filename":"Onlinefloatimage2.png","url":"https://assets-eu.researchsquare.com/files/rs-7367912/v1/b9b7cb8b38be2ad52008851f.png"},{"id":92023982,"identity":"c0f797a6-3f84-4bcf-ac57-0d080cb30411","added_by":"auto","created_at":"2025-09-23 18:27:14","extension":"png","order_by":38,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":2150,"visible":true,"origin":"","legend":"","description":"","filename":"Onlinefloatimage20.png","url":"https://assets-eu.researchsquare.com/files/rs-7367912/v1/d7300b27b043b9604dcd4e1a.png"},{"id":92023525,"identity":"db573d23-0e8d-48c0-9681-ce583935398b","added_by":"auto","created_at":"2025-09-23 18:19:14","extension":"png","order_by":39,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":2796,"visible":true,"origin":"","legend":"","description":"","filename":"Onlinefloatimage21.png","url":"https://assets-eu.researchsquare.com/files/rs-7367912/v1/7bc8d43c8d193944d345391a.png"},{"id":92023526,"identity":"a022a9a4-8280-4df8-981e-42e696a06341","added_by":"auto","created_at":"2025-09-23 18:19:14","extension":"png","order_by":40,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":2152,"visible":true,"origin":"","legend":"","description":"","filename":"Onlinefloatimage22.png","url":"https://assets-eu.researchsquare.com/files/rs-7367912/v1/e7171a385ddb21cffb950d50.png"},{"id":92023985,"identity":"c2586ebe-1158-42f8-a4c5-31ffe7a3d559","added_by":"auto","created_at":"2025-09-23 18:27:14","extension":"png","order_by":41,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":2928,"visible":true,"origin":"","legend":"","description":"","filename":"Onlinefloatimage23.png","url":"https://assets-eu.researchsquare.com/files/rs-7367912/v1/072a2e9490f9e76e214144d9.png"},{"id":92023162,"identity":"5091b7ce-98fb-4ca8-a802-97a8324a5557","added_by":"auto","created_at":"2025-09-23 18:11:14","extension":"png","order_by":42,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":68557,"visible":true,"origin":"","legend":"","description":"","filename":"Onlinefloatimage3.png","url":"https://assets-eu.researchsquare.com/files/rs-7367912/v1/aa2cbf825a3efc829d663232.png"},{"id":92023528,"identity":"fdc88c80-7b2b-430b-a4c9-409491ec7751","added_by":"auto","created_at":"2025-09-23 18:19:14","extension":"png","order_by":43,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":29622,"visible":true,"origin":"","legend":"","description":"","filename":"Onlinefloatimage4.png","url":"https://assets-eu.researchsquare.com/files/rs-7367912/v1/7237dc2494e6e0b87babbf5a.png"},{"id":92023984,"identity":"1c9d66d9-8e16-438d-8d31-1ed47bc1ba66","added_by":"auto","created_at":"2025-09-23 18:27:14","extension":"png","order_by":44,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":185692,"visible":true,"origin":"","legend":"","description":"","filename":"Onlinefloatimage5.png","url":"https://assets-eu.researchsquare.com/files/rs-7367912/v1/665fd4d14abb3700fb2346b0.png"},{"id":92023517,"identity":"7b2a7c54-6823-45bd-947b-f31866cc1802","added_by":"auto","created_at":"2025-09-23 18:19:14","extension":"png","order_by":45,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":21730,"visible":true,"origin":"","legend":"","description":"","filename":"Onlinefloatimage6.png","url":"https://assets-eu.researchsquare.com/files/rs-7367912/v1/d77b068176ab916946a90bcb.png"},{"id":92023176,"identity":"207bf925-ce68-4317-8a70-416625486aca","added_by":"auto","created_at":"2025-09-23 18:11:14","extension":"png","order_by":46,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":26666,"visible":true,"origin":"","legend":"","description":"","filename":"Onlinefloatimage7.png","url":"https://assets-eu.researchsquare.com/files/rs-7367912/v1/a8e0f884c9edb2569d4dee21.png"},{"id":92023522,"identity":"5b77142a-05c2-4b59-be6f-792e737b0073","added_by":"auto","created_at":"2025-09-23 18:19:14","extension":"png","order_by":47,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":1837,"visible":true,"origin":"","legend":"","description":"","filename":"Onlinefloatimage8.png","url":"https://assets-eu.researchsquare.com/files/rs-7367912/v1/437253608b05f9d3ad828de7.png"},{"id":92023172,"identity":"1024555b-047f-40e1-86f0-8de89bd731b3","added_by":"auto","created_at":"2025-09-23 18:11:14","extension":"png","order_by":48,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":1680,"visible":true,"origin":"","legend":"","description":"","filename":"Onlinefloatimage9.png","url":"https://assets-eu.researchsquare.com/files/rs-7367912/v1/837886104e6c00984d6ca5c5.png"},{"id":92023986,"identity":"0e01e341-fb3c-405f-9a7d-2f17f6b44be1","added_by":"auto","created_at":"2025-09-23 18:27:14","extension":"xml","order_by":49,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":119247,"visible":true,"origin":"","legend":"","description":"","filename":"2156da1fbb8e45b29ea9f785f69b721e1structuring.xml","url":"https://assets-eu.researchsquare.com/files/rs-7367912/v1/f27c509af163798c26104af2.xml"},{"id":92023181,"identity":"71c45905-edd8-4027-a032-9d4bfbb94714","added_by":"auto","created_at":"2025-09-23 18:11:14","extension":"html","order_by":50,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":133250,"visible":true,"origin":"","legend":"","description":"","filename":"earlyproof.html","url":"https://assets-eu.researchsquare.com/files/rs-7367912/v1/81f73da79ecfa09be29dfb23.html"},{"id":92023124,"identity":"7f205276-cb5c-4cbd-b188-5420a491f888","added_by":"auto","created_at":"2025-09-23 18:11:13","extension":"png","order_by":1,"title":"Figure 1","display":"","copyAsset":false,"role":"figure","size":150966,"visible":true,"origin":"","legend":"\u003cp\u003eExamples of Arabic CAPTCHA images generated using the proposed Pseudoisochromatic Plate-based approach, incorporating both printed and handwritten text styles\u003c/p\u003e","description":"","filename":"1.png","url":"https://assets-eu.researchsquare.com/files/rs-7367912/v1/bc1f28d4a652e357e18ece70.png"},{"id":92023125,"identity":"f17b0f65-319d-427c-9834-92191da6ab02","added_by":"auto","created_at":"2025-09-23 18:11:13","extension":"png","order_by":2,"title":"Figure 2","display":"","copyAsset":false,"role":"figure","size":239597,"visible":true,"origin":"","legend":"\u003cp\u003e(a) The Stilling test, (b) The Ishihara test, (c) The Hardy test, (d) The Cambridge Colour Test\u003c/p\u003e","description":"","filename":"2.png","url":"https://assets-eu.researchsquare.com/files/rs-7367912/v1/810aedfa495be97799045082.png"},{"id":92023974,"identity":"04c41789-d66a-4fe3-b5dd-8f1a0fab7594","added_by":"auto","created_at":"2025-09-23 18:27:13","extension":"png","order_by":3,"title":"Figure 3","display":"","copyAsset":false,"role":"figure","size":531211,"visible":true,"origin":"","legend":"\u003cp\u003eSchematic Representation of the Proposed CAPTCHA Framework\u003c/p\u003e","description":"","filename":"3.png","url":"https://assets-eu.researchsquare.com/files/rs-7367912/v1/ddea744a2619082e03a4a188.png"},{"id":92023131,"identity":"a466a677-9778-4dd8-9cec-be13e9646ac7","added_by":"auto","created_at":"2025-09-23 18:11:13","extension":"png","order_by":4,"title":"Figure 4","display":"","copyAsset":false,"role":"figure","size":16689,"visible":true,"origin":"","legend":"\u003cp\u003eExamples of generated printed Arabic text samples (meaningful (a) and meaningless(b))\u003c/p\u003e","description":"","filename":"4.png","url":"https://assets-eu.researchsquare.com/files/rs-7367912/v1/9628a7e098f6d295573057e3.png"},{"id":92023127,"identity":"0cf9eec4-e78d-415a-918f-54f09e5f43a0","added_by":"auto","created_at":"2025-09-23 18:11:13","extension":"png","order_by":5,"title":"Figure 5","display":"","copyAsset":false,"role":"figure","size":30540,"visible":true,"origin":"","legend":"\u003cp\u003eExamples of generated handwritten Arabic text samples (meaningful (a) and meaningless(b))\u003c/p\u003e","description":"","filename":"5.png","url":"https://assets-eu.researchsquare.com/files/rs-7367912/v1/2f8905015ca0aeff2329db79.png"},{"id":92023502,"identity":"a833aeca-3c3a-49b3-bbef-a3b10c308a11","added_by":"auto","created_at":"2025-09-23 18:19:13","extension":"png","order_by":6,"title":"Figure 6","display":"","copyAsset":false,"role":"figure","size":1103143,"visible":true,"origin":"","legend":"\u003cp\u003eFinal appearance of the generated CAPTCHA samples previously shown in Figure 4 — (a) meaningful printed text and (b) meaningless printed text — and Figure 5 — (c) meaningful handwritten text and (d) meaningless handwritten text\u003c/p\u003e","description":"","filename":"6.png","url":"https://assets-eu.researchsquare.com/files/rs-7367912/v1/468e4d2e6419c30a1c40da2a.png"},{"id":92023976,"identity":"ce312229-e990-4455-80e3-6834c4be8b0e","added_by":"auto","created_at":"2025-09-23 18:27:13","extension":"png","order_by":7,"title":"Figure 7","display":"","copyAsset":false,"role":"figure","size":248434,"visible":true,"origin":"","legend":"\u003cp\u003eA sample of implementing distractor types parameter\u003c/p\u003e","description":"","filename":"7.png","url":"https://assets-eu.researchsquare.com/files/rs-7367912/v1/37d4c4a0075e88dabbfc4714.png"},{"id":92023500,"identity":"2e1bbf48-4185-403b-b0d7-85cf59b4039f","added_by":"auto","created_at":"2025-09-23 18:19:13","extension":"png","order_by":8,"title":"Figure 8","display":"","copyAsset":false,"role":"figure","size":100354,"visible":true,"origin":"","legend":"\u003cp\u003eEfficiency comparison between the Baseline and Proposed CAPTCHA schemes across four categories: handwritten meaningful, handwritten meaningless, printed meaningful, and printed meaningless.\u003c/p\u003e","description":"","filename":"8.png","url":"https://assets-eu.researchsquare.com/files/rs-7367912/v1/5c17e602fcb7a4b93bb5d64b.png"},{"id":92023140,"identity":"4dc175a4-516a-41e0-bd0c-e9e5647e7f96","added_by":"auto","created_at":"2025-09-23 18:11:13","extension":"png","order_by":9,"title":"Figure 9","display":"","copyAsset":false,"role":"figure","size":92801,"visible":true,"origin":"","legend":"\u003cp\u003eComparison of CAPTCHA effectiveness between the Baseline and Proposed schemes across four categories: handwritten meaningful, handwritten meaningless, printed meaningful, and printed meaningless.\u003c/p\u003e","description":"","filename":"9.png","url":"https://assets-eu.researchsquare.com/files/rs-7367912/v1/78f2e44f997691e97e790e06.png"},{"id":92023146,"identity":"aef0d658-4259-42a8-8ad6-174ee3b37a5b","added_by":"auto","created_at":"2025-09-23 18:11:14","extension":"png","order_by":10,"title":"Figure 10","display":"","copyAsset":false,"role":"figure","size":583848,"visible":true,"origin":"","legend":"\u003cp\u003eThe results of satisfaction survey\u003c/p\u003e","description":"","filename":"10.png","url":"https://assets-eu.researchsquare.com/files/rs-7367912/v1/519aa90b9330fba0d6cf63e2.png"},{"id":93003697,"identity":"7da48e10-db10-4337-8b46-f46a39ac2b87","added_by":"auto","created_at":"2025-10-08 06:10:08","extension":"pdf","order_by":0,"title":"","display":"","copyAsset":false,"role":"manuscript-pdf","size":4096127,"visible":true,"origin":"","legend":"","description":"","filename":"manuscript.pdf","url":"https://assets-eu.researchsquare.com/files/rs-7367912/v1/c15f3cbc-4460-4a1b-8f97-03af4a31437d.pdf"}],"financialInterests":"No competing interests reported.","formattedTitle":"A Novel Arabic CAPTCHA Generation Framework Based on Pseudoisochromatic Plates for Enhanced Human Verification","fulltext":[{"header":"1. Introduction","content":"\u003cdiv class=\"BlockQuote\"\u003e\n\u003cp\u003eThe widespread reliance on online services has intensified the need for secure and user-friendly human verification mechanisms. Completely Automated Public Turing tests to tell Computers and Humans Apart (CAPTCHAs) are widely deployed to prevent automated bots from abusing web services [\u003cspan class=\"CitationRef\"\u003e1\u003c/span\u003e]. While traditional CAPTCHAs, such as distorted text, image selection, or click-based puzzles, have proven effective to some extent, they are increasingly vulnerable to advanced machine learning and sophisticated optical character recognition (OCR) techniques (e.g, [\u003cspan class=\"CitationRef\"\u003e2\u003c/span\u003e]). Moreover, CAPTCHAs should also balance security with usability to avoid frustrating legitimate users [\u003cspan class=\"CitationRef\"\u003e3\u003c/span\u003e].\u003c/p\u003e\n\u003cp\u003eArabic, as one of the most widely spoken languages globally and a primary language in many digital markets, presents unique challenges in the context of CAPTCHA design. The cursive nature of Arabic script, its contextual letter shapes, and diacritical marks pose difficulties for both OCR systems and CAPTCHA generators. The groundbreaking papers on Arabic CAPTCHAs were [\u003cspan class=\"CitationRef\"\u003e4\u003c/span\u003e\u0026ndash;\u003cspan class=\"CitationRef\"\u003e5\u003c/span\u003e] as Arabic letters were shown in the CAPTCHAs. These were followed by a set of studies (e.g., [\u003cspan class=\"CitationRef\"\u003e6\u003c/span\u003e\u0026ndash;\u003cspan class=\"CitationRef\"\u003e9\u003c/span\u003e]) that used both Arabic printed and handwritten letters. The output of these studies showed promising results. Despite these results, most existing Arabic CAPTCHA schemes are either direct adaptations of English models or rely on distorted handwritten fonts, making them susceptible to pattern recognition-based attacks or difficult for users to solve. There is a pressing need for CAPTCHA approaches that are inherently more secure, linguistically appropriate, and accessible to a diverse population.\u003c/p\u003e\n\u003cp\u003eThus, this paper proposes a novel Arabic CAPTCHA generation method based on \u003cstrong\u003ePseudoisochromatic Plates (PIPs)\u003c/strong\u003e\u0026mdash;a visual technique originally developed for diagnosing color vision deficiencies [\u003cspan class=\"CitationRef\"\u003e10\u003c/span\u003e]. PIPs consist of color-blended backgrounds embedded with readable characters or shapes discernible only to individuals with normal color perception. By adapting this concept, we introduce a CAPTCHA model that leverages subtle color variations and visual noise to obscure Arabic text in a way that is easily interpreted by humans but yet challenging for automated solvers. Unlike conventional noise-based or distorted-text CAPTCHAs [\u003cspan class=\"CitationRef\"\u003e11\u003c/span\u003e], PIP-inspired CAPTCHAs exploit the complexity of human color perception, thereby offering a new layer of resistance against machine-based attacks.\u003c/p\u003e\n\u003cp\u003eAs an illustration of the proposed approach, Fig.\u0026nbsp;1 presents representative samples generated using our developed generator. Specifically, (a) displays a sample utilizing printed Arabic text, while (b) shows a counterpart rendered in handwritten style. Although recent studies (e.g., [\u003cspan class=\"CitationRef\"\u003e11\u003c/span\u003e][\u003cspan class=\"CitationRef\"\u003e12\u003c/span\u003e]) have emphasized the superior resistance of handwritten Arabic CAPTCHAs to automated attacks, incorporating the printed text style within this framework allows for a comparative investigation into the security and usability trade-offs\u0026mdash;thereby contributing further insights to the existing literature. This dual-style integration not only broadens the applicability of the proposed CAPTCHA system across diverse user groups, but also facilitates a comprehensive evaluation of human interaction dynamics and machine-solving resistance. By juxtaposing both styles within a unified generation framework, the proposed method enables a deeper understanding of how visual complexity, font variability, and contextual legibility influence CAPTCHA effectiveness. Furthermore, the use of the PIPs technique enhances visual noise and perceptual ambiguity\u0026mdash;essential features for countering automated recognition tools\u0026mdash;while maintaining readability for human users. This balance is critical for designing CAPTCHAs that are both user-friendly and resilient against adversarial machine-learning models.\u003c/p\u003e\n\u003c/div\u003e\n\u003cp\u003eThe proposed CAPTCHA scheme was evaluated in terms of both usability and security through a series of experimental studies. The results demonstrate that human users were able to solve the CAPTCHA with significantly higher effectiveness compared to machine-based algorithms. Furthermore, the findings indicate strong resistance to automated attacks, highlighting the robustness of the proposed scheme.\u003c/p\u003e\n\u003cp\u003eThe proposed approach contributes to the field in several ways. First, it innovatively integrates PIP principles into CAPTCHA design, enhancing security through perceptual obfuscation rather than distortion alone. Second, the approach is evaluated through a series of experiments involving human participants and automated attack models to assess both usability and robustness. Finally, it categorizes exist studies based on the type of text employed. Thus, this study explores a novel direction in CAPTCHA development by fusing Arabic script characteristics with the perceptual challenges of PIPs. The findings aim to advance the design of CAPTCHAs that are both secure and inclusive, particularly in Arabic-speaking and visually diverse user communities.\u003c/p\u003e\n\u003cp\u003eThe remainder of this paper is organized as follows: Section 2 reviews the related work. Section 3 provides an overview of the PIPs test. Section 4 outlines the proposed methodology, while Section 5 describes the experimental setup and evaluation. Section 6 presents the results, which are further discussed in Section 7. Finally, Section 8 concludes the paper.\u003c/p\u003e"},{"header":"2. Related Work","content":"\u003cp\u003e\u003cdiv class=\"BlockQuote\"\u003e\u003cp\u003eTo the best of our knowledge, this paper is the first work for exploiting PIP approach to be applied as text-based Arabic CAPTCHAs. Therefore, this section shows a comprehensive review of the state of the art of Arabic CAPTCHA schemes.\u003c/p\u003e\u003cp\u003eArabic printed text-based CAPTCHAs have been widely explored for enhancing online security. Shirali-Shahreza and Shirali-Shahreza introduced \"Baffletext\" CAPTCHA for Persian and Arabic scripts, emphasizing text distortion to counter OCR attacks [\u003cspan citationid=\"CR4\" class=\"CitationRef\"\u003e4\u003c/span\u003e]. Shahreza applied Arabic CAPTCHA specifically to spam SMS verification [\u003cspan citationid=\"CR5\" class=\"CitationRef\"\u003e5\u003c/span\u003e], while Khan et al. evaluated its effectiveness against automated intrusions [\u003cspan citationid=\"CR6\" class=\"CitationRef\"\u003e6\u003c/span\u003e, \u003cspan citationid=\"CR13\" class=\"CitationRef\"\u003e13\u003c/span\u003e]. Banday and Sheikh reviewed various CAPTCHA types, highlighting Arabic variants [\u003cspan citationid=\"CR14\" class=\"CitationRef\"\u003e14\u003c/span\u003e]. To increase unpredictability, Sulaiman and Hassan proposed random Arabic character generation [\u003cspan citationid=\"CR15\" class=\"CitationRef\"\u003e15\u003c/span\u003e], and Alsuhibany et al. analyzed difficulty and robustness against OCR and segmentation attacks [\u003cspan citationid=\"CR16\" class=\"CitationRef\"\u003e16\u003c/span\u003e, \u003cspan citationid=\"CR18\" class=\"CitationRef\"\u003e18\u003c/span\u003e]. Sulaiman also explored chaotic maps to introduce unpredictability [\u003cspan citationid=\"CR17\" class=\"CitationRef\"\u003e17\u003c/span\u003e]. Moreover, multilingual [\u003cspan citationid=\"CR19\" class=\"CitationRef\"\u003e19\u003c/span\u003e] and stylistically complex Nastaliq script-based CAPTCHAs were introduced to further enhance security [\u003cspan citationid=\"CR20\" class=\"CitationRef\"\u003e20\u003c/span\u003e]. Collectively, these studies underscore the importance of linguistic complexity, randomness, and stylistic diversity in strengthening Arabic CAPTCHA systems against automated attacks.\u003c/p\u003e\u003cp\u003eRecent research has explored various Arabic CAPTCHA approaches, emphasizing security through handwritten and multilingual methods. Alsuhibany and Parvez [\u003cspan citationid=\"CR1\" class=\"CitationRef\"\u003e1\u003c/span\u003e] developed a secure Arabic handwritten CAPTCHA resistant to OCR attacks, while Alrasheed and Alsuhibany [\u003cspan citationid=\"CR3\" class=\"CitationRef\"\u003e3\u003c/span\u003e] employed adversarial techniques to further enhance security. Alsuhibany and Alquraishi [\u003cspan citationid=\"CR21\" class=\"CitationRef\"\u003e21\u003c/span\u003e] explored Arabic CAPTCHA usability via visual cryptography, and Parvez et al. [\u003cspan citationid=\"CR22\" class=\"CitationRef\"\u003e22\u003c/span\u003e] introduced an Arabic CAPTCHA gamification framework for cybersecurity education. Lajmi et al. [\u003cspan citationid=\"CR23\" class=\"CitationRef\"\u003e23\u003c/span\u003e] strengthened security through calligraphy-based handwritten CAPTCHAs, whereas Aldosari [\u003cspan citationid=\"CR24\" class=\"CitationRef\"\u003e24\u003c/span\u003e, \u003cspan citationid=\"CR26\" class=\"CitationRef\"\u003e26\u003c/span\u003e] presented multilingual handwritten CAPTCHA methods. Additionally, Parvez and Alsuhibany [\u003cspan citationid=\"CR25\" class=\"CitationRef\"\u003e25\u003c/span\u003e] proposed a segmentation-validation technique to generate robust handwritten Arabic CAPTCHAs. Collectively, these studies demonstrate the effectiveness of handwriting styles, multilingualism, and advanced validation techniques in enhancing Arabic CAPTCHA security and usability.\u003c/p\u003e\u003cp\u003eFawa\u0026rsquo;reh et al. [\u003cspan citationid=\"CR27\" class=\"CitationRef\"\u003e27\u003c/span\u003e] proposed an Arabic CAPTCHA scheme designed to mitigate deep learning-based attacks targeting text image CAPTCHAs. Their approach introduces unique visual distortions and features tailored to the Arabic script, enhancing resistance to automated recognition methods commonly employed in adversarial machine learning. The study highlights the effectiveness of culturally adapted CAPTCHA designs in improving security against sophisticated AI-driven threats.\u003c/p\u003e\u003cp\u003eRecent research has focused on enhancing the interactivity and security of Arabic CAPTCHAs. Alsuhibany and Parvez [\u003cspan citationid=\"CR12\" class=\"CitationRef\"\u003e12\u003c/span\u003e] proposed attack-filtered interactive Arabic CAPTCHAs that adapt based on detected attack patterns, aiming to increase robustness against automated solvers. In a comparative study, Alsuhibany and Alnoshan [\u003cspan citationid=\"CR28\" class=\"CitationRef\"\u003e28\u003c/span\u003e] evaluated interactive handwritten and text-based handwritten Arabic CAPTCHA schemes tailored for mobile devices, highlighting trade-offs between usability and resilience. Additionally, Alsuhibany and Alquraishi [\u003cspan citationid=\"CR21\" class=\"CitationRef\"\u003e21\u003c/span\u003e] employed visual cryptography to improve the security and usability of Arabic text-based CAPTCHAs, offering a novel user-verifiable authentication mechanism.\u003c/p\u003e\u003cp\u003eSeveral studies have explored Arabic reCAPTCHA systems aimed at improving web security and digitizing Arabic content. Bakry et al. [\u003cspan citationid=\"CR8\" class=\"CitationRef\"\u003e8\u003c/span\u003e] introduced ARECAPTCHA, which leverages crowdsourcing from native speakers to digitize Arabic text. Building on this, Abubaker et al. proposed a cloud-based Arabic reCAPTCHA service, detailing its design and architecture across multiple works [\u003cspan citationid=\"CR9\" class=\"CitationRef\"\u003e9\u003c/span\u003e, \u003cspan citationid=\"CR10\" class=\"CitationRef\"\u003e10\u003c/span\u003e], and emphasizing its application in enhancing the digitization of Arabic manuscripts [\u003cspan citationid=\"CR30\" class=\"CitationRef\"\u003e30\u003c/span\u003e, \u003cspan citationid=\"CR31\" class=\"CitationRef\"\u003e31\u003c/span\u003e]. Their research also addressed the broader challenges and opportunities of Arabic digital content [\u003cspan citationid=\"CR29\" class=\"CitationRef\"\u003e29\u003c/span\u003e], presenting the reCAPTCHA system as a dual-purpose tool for both security and heritage preservation.\u003c/p\u003e\u003cp\u003eTo provide a clearer understanding of the current landscape of Arabic CAPTCHA research, the reviewed studies are summarized in Table\u0026nbsp;\u003cspan refid=\"Tab1\" class=\"InternalRef\"\u003e1\u003c/span\u003e. This table categorizes each study based on the type of text employed\u0026mdash;whether meaningful (semantically valid Arabic words or phrases), meaningless (random or distorted characters), or individual characters. Such classification helps highlight trends in usability, readability, and security across different CAPTCHA designs.\u003c/p\u003e\u003c/div\u003e\u003c/p\u003e\u003cp\u003e\u003cdiv class=\"gridtable\"\u003e\u003ctable float=\"Yes\" id=\"Tab1\" border=\"1\"\u003e\u003ccaption language=\"En\"\u003e\u003cdiv class=\"CaptionNumber\"\u003eTable 1\u003c/div\u003e\u003cdiv class=\"CaptionContent\"\u003e\u003cp\u003eSummary of Reviewed Arabic CAPTCHA Studies by Text Type (Meaningful vs. Meaningless)\u003c/p\u003e\u003c/div\u003e\u003c/caption\u003e\u003ccolgroup cols=\"4\"\u003e\u003cdiv align=\"left\" class=\"colspec\" colname=\"c1\" colnum=\"1\"\u003e\u003c/div\u003e\u003cdiv align=\"left\" class=\"colspec\" colname=\"c2\" colnum=\"2\"\u003e\u003c/div\u003e\u003cdiv align=\"left\" class=\"colspec\" colname=\"c3\" colnum=\"3\"\u003e\u003c/div\u003e\u003cdiv align=\"left\" class=\"colspec\" colname=\"c4\" colnum=\"4\"\u003e\u003c/div\u003e\u003cthead\u003e\u003ctr\u003e\u003cth align=\"left\" colname=\"c1\" morerows=\"1\" rowspan=\"2\"\u003e\u003cp\u003eType of Arabic CAPTCHAs\u003c/p\u003e\u003c/th\u003e\u003cth align=\"left\" colspan=\"3\" nameend=\"c4\" namest=\"c2\"\u003e\u003cp\u003eStudy\u003c/p\u003e\u003c/th\u003e\u003c/tr\u003e\u003ctr\u003e\u003cth align=\"left\" colname=\"c2\"\u003e\u003cp\u003eMeaningless\u003c/p\u003e\u003c/th\u003e\u003cth align=\"left\" colname=\"c3\"\u003e\u003cp\u003eMeaningful\u003c/p\u003e\u003c/th\u003e\u003cth align=\"left\" colname=\"c4\"\u003e\u003cp\u003eCharacters\u003c/p\u003e\u003c/th\u003e\u003c/tr\u003e\u003c/thead\u003e\u003ctbody\u003e\u003ctr\u003e\u003ctd align=\"left\" colname=\"c1\"\u003e\u003cp\u003ePrinted text-based\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c2\"\u003e\u003cp\u003e[\u003cspan additionalcitationids=\"CR5\" citationid=\"CR4\" class=\"CitationRef\"\u003e4\u003c/span\u003e\u0026ndash;\u003cspan citationid=\"CR6\" class=\"CitationRef\"\u003e6\u003c/span\u003e] [\u003cspan additionalcitationids=\"CR14\" citationid=\"CR13\" class=\"CitationRef\"\u003e13\u003c/span\u003e\u0026ndash;\u003cspan citationid=\"CR15\" class=\"CitationRef\"\u003e15\u003c/span\u003e][\u003cspan citationid=\"CR17\" class=\"CitationRef\"\u003e17\u003c/span\u003e]\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c3\"\u003e\u003cp\u003e[\u003cspan citationid=\"CR7\" class=\"CitationRef\"\u003e7\u003c/span\u003e][\u003cspan citationid=\"CR16\" class=\"CitationRef\"\u003e16\u003c/span\u003e][\u003cspan citationid=\"CR18\" class=\"CitationRef\"\u003e18\u003c/span\u003e]\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c4\"\u003e\u003cp\u003e[\u003cspan citationid=\"CR19\" class=\"CitationRef\"\u003e19\u003c/span\u003e\u0026ndash;\u003cspan citationid=\"CR20\" class=\"CitationRef\"\u003e20\u003c/span\u003e]\u003c/p\u003e\u003c/td\u003e\u003c/tr\u003e\u003ctr\u003e\u003ctd align=\"left\" colname=\"c1\"\u003e\u003cp\u003eHandwritten text-based\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c2\"\u003e\u003cp\u003e[\u003cspan additionalcitationids=\"CR25\" citationid=\"CR24\" class=\"CitationRef\"\u003e24\u003c/span\u003e\u0026ndash;\u003cspan citationid=\"CR26\" class=\"CitationRef\"\u003e26\u003c/span\u003e]\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c3\"\u003e\u003cp\u003e[\u003cspan citationid=\"CR1\" class=\"CitationRef\"\u003e1\u003c/span\u003e][\u003cspan citationid=\"CR3\" class=\"CitationRef\"\u003e3\u003c/span\u003e][\u003cspan citationid=\"CR21\" class=\"CitationRef\"\u003e21\u003c/span\u003e\u0026ndash;\u003cspan citationid=\"CR22\" class=\"CitationRef\"\u003e22\u003c/span\u003e]\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c4\"\u003e\u003cp\u003e[\u003cspan citationid=\"CR23\" class=\"CitationRef\"\u003e23\u003c/span\u003e]\u003c/p\u003e\u003c/td\u003e\u003c/tr\u003e\u003ctr\u003e\u003ctd align=\"left\" colname=\"c1\"\u003e\u003cp\u003eImage-based\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c2\"\u003e\u003cp\u003e-\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c3\"\u003e\u003cp\u003e-\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c4\"\u003e\u003cp\u003e[\u003cspan citationid=\"CR27\" class=\"CitationRef\"\u003e27\u003c/span\u003e]\u003c/p\u003e\u003c/td\u003e\u003c/tr\u003e\u003ctr\u003e\u003ctd align=\"left\" colname=\"c1\"\u003e\u003cp\u003eInteractive\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c2\"\u003e\u003cp\u003e[\u003cspan citationid=\"CR12\" class=\"CitationRef\"\u003e12\u003c/span\u003e][\u003cspan citationid=\"CR28\" class=\"CitationRef\"\u003e28\u003c/span\u003e]\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c3\"\u003e\u003cp\u003e[\u003cspan citationid=\"CR21\" class=\"CitationRef\"\u003e21\u003c/span\u003e]\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c4\"\u003e\u003cp\u003e-\u003c/p\u003e\u003c/td\u003e\u003c/tr\u003e\u003ctr\u003e\u003ctd align=\"left\" colname=\"c1\"\u003e\u003cp\u003ereCAPTCHA\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c2\"\u003e\u003cp\u003e-\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c3\"\u003e\u003cp\u003e[\u003cspan additionalcitationids=\"CR9\" citationid=\"CR8\" class=\"CitationRef\"\u003e8\u003c/span\u003e\u0026ndash;\u003cspan citationid=\"CR10\" class=\"CitationRef\"\u003e10\u003c/span\u003e][\u003cspan citationid=\"CR29\" class=\"CitationRef\"\u003e29\u003c/span\u003e][\u003cspan citationid=\"CR30\" class=\"CitationRef\"\u003e30\u003c/span\u003e][\u003cspan citationid=\"CR31\" class=\"CitationRef\"\u003e31\u003c/span\u003e]\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c4\"\u003e\u0026nbsp;\u003c/td\u003e\u003c/tr\u003e\u003c/tbody\u003e\u003c/colgroup\u003e\u003c/table\u003e\u003c/div\u003e\u003c/p\u003e"},{"header":"3. Pseudoisochromatic Plate Test: An Overview","content":"\u003cdiv\u003e\n\u003cp\u003ePIP tests are widely recognized as standard tools for screening, diagnosing, and classifying color vision deficiencies (CVD) [32]. These tests typically consist of colored dot patterns that form identifiable figures\u0026mdash;such as numbers, letters, lines, or geometric shapes\u0026mdash;that are distinguishable from the background only through color differences. The visibility of these figures depends on chromatic contrasts; for individuals with CVD, such contrast may be imperceptible, rendering the figure invisible or misidentified [33].\u003c/p\u003e\n\u003cp\u003eThe effectiveness of these tests is rooted in the opponent-process theory of color vision, which posits that visual perception is governed by three antagonistic color pairs: black-white, red-green, and blue-yellow [34]. In individuals with color vision deficiency, two or more colors from an opponent pair may appear identical, resulting in isochromatic perception where the figure becomes indistinguishable from its background.\u003c/p\u003e\n\u003cp\u003eThe concept of PIP testing originated with Jakob Stilling, who developed the first clinical color vision test using this principle [32]. Although Stilling\u0026rsquo;s test was groundbreaking, it had several limitations, which were later addressed by the development of the \u003cstrong\u003eIshihara plates\u003c/strong\u003e, now among the most widely used PIP tests worldwide [35]. Another well-known variant is the \u003cstrong\u003eHardy-Rand-Rittler (HRR)\u003c/strong\u003e test, which uses geometric symbols like circles, crosses, and triangles in a specific chromatic arrangement to challenge color-deficient observers [36].\u003c/p\u003e\n\u003cp\u003eIn recent years, digital implementations of PIP tests have emerged to improve standardization and repeatability. Computerized systems, such as the \u003cstrong\u003eCambridge Colour Test (CCT)\u003c/strong\u003e, offer adaptive difficulty and precise control over chromatic parameters. The CCT presents a colored Landholt C that must be identified based solely on chromatic differences, enabling detailed measurement of color discrimination thresholds [37]. Figure\u0026nbsp;2 illustrates a sample for aforementioned test types.\u003c/p\u003e\n\u003c/div\u003e\n\u003cdiv\u003e\u0026nbsp;\u003c/div\u003e\n\u003cp\u003eOverall, PIP tests remain indispensable in both clinical and research settings due to their simplicity, non-invasiveness, and diagnostic value. The transition from static printed plates to digital, interactive formats underscores their adaptability to modern needs. This versatility opens new avenues for novel applications\u0026mdash;one of which is the generation of visual challenges in CAPTCHA systems. Specifically, the inherent difficulty PIP designs pose for automated systems, while remaining interpretable by humans with normal color vision, makes them particularly well-suited for developing secure, user-friendly Arabic CAPTCHAs. By embedding Arabic text within color-differentiated figure-background patterns, akin to those used in PIP tests, it is possible to create CAPTCHA schemes that are resistant to machine learning-based attacks while maintaining high usability for native Arabic speakers. However, since PIP tests are widely recognized as standard tools for screening, diagnosing, and classifying color vision deficiencies, careful refinement will be necessary to ensure that the proposed approach is accessible to all users\u0026mdash;not only those with normal color vision. This consideration will be further discussed in the following section.\u003c/p\u003e"},{"header":"4. Methodology","content":"\u003cdiv class=\"BlockQuote\"\u003e\n\u003cp\u003eThis section outlines the design and evaluation of our PIP\u0026ndash;based Arabic CAPTCHA system. We first describe the sample generation process for both printed and handwritten text styles using PIPs. We then present our testing procedures\u0026mdash;covering adversarial resilience, usability assessments, and evaluation metrics\u0026mdash;to demonstrate the framework\u0026rsquo;s security and user-friendliness.\u003c/p\u003e\n\u003c/div\u003e\n\u003cdiv id=\"Sec5\" class=\"Section2\"\u003e\n\u003ch2\u003e4.1. Framework of the Proposed CAPTCHA Scheme\u003c/h2\u003e\n\u003cdiv class=\"BlockQuote\"\u003e\n\u003cp\u003eThis section introduces the proposed CAPTCHA design framework, which combines Arabic text generation with a PIP-style background to develop a secure and user-accessible challenge mechanism, as shown in Fig.\u0026nbsp;\u003cspan class=\"InternalRef\"\u003e3\u003c/span\u003e. The framework follows a structured pipeline that begins with the generation of Arabic text or words, followed by an embedding phase in which the text is integrated into a PIP-like background using a multistage color blending process. This embedding stage consists of three key phases: \u003cstrong\u003ecolor selection\u003c/strong\u003e, \u003cstrong\u003eword representation\u003c/strong\u003e, and \u003cstrong\u003eimage rendering\u003c/strong\u003e\u0026mdash;each contributing to the visual complexity and robustness of the CAPTCHA.\u003c/p\u003e\n\u003cp\u003eTo ensure comprehensive evaluation, the framework also incorporates two core considerations: \u003cstrong\u003esecurity\u003c/strong\u003e and \u003cstrong\u003eusability\u003c/strong\u003e. The security dimension focuses on resistance to current automated recognition techniques and challenges posed by color-blind bots, while the usability dimension emphasizes human readability, accessibility across diverse visual capabilities, and suitability for various user groups. These phases are discussed in the following sections.\u003c/p\u003e\n\u003c/div\u003e\n\u003cp\u003e\u0026nbsp;\u003c/p\u003e\n\u003cdiv id=\"Sec6\" class=\"Section3\"\u003e\n\u003ch2\u003e4.1.1. Generation of Arabic text/word\u003c/h2\u003e\n\u003cdiv class=\"BlockQuote\"\u003e\n\u003cp\u003eThis subsection outlines the procedures used to generate the Arabic text that forms the core content of the proposed CAPTCHA scheme. Two distinct types of Arabic text generation are considered: \u003cstrong\u003eprinted\u003c/strong\u003e and \u003cstrong\u003ehandwritten\u003c/strong\u003e. Each type serves a unique role in assessing the system\u0026rsquo;s robustness and usability. To facilitate structured implementation and analysis, the generation process is divided into two categories: (1) \u003cstrong\u003ePrinted Arabic text generation\u003c/strong\u003e, and (2) \u003cstrong\u003eHandwritten Arabic text generation\u003c/strong\u003e. Each category employs its own dedicated generation method tailored to its respective style and application context.\u003c/p\u003e\n\u003cp\u003eAccordingly, we employed the text-generation tool described in [\u003cspan class=\"CitationRef\"\u003e21\u003c/span\u003e] to produce both meaningful and meaningless printed Arabic samples. Representative outputs are shown in Fig.\u0026nbsp;4.\u003c/p\u003e\n\u003c/div\u003e\n\u003cp\u003eMeaningless handwritten samples were produced using the text-generation tool described in [\u003cspan class=\"CitationRef\"\u003e38\u003c/span\u003e], while meaningful handwritten words were generated according to the approach outlined in [\u003cspan class=\"CitationRef\"\u003e3\u003c/span\u003e]. Representative outputs are shown in Fig.\u0026nbsp;5.\u003c/p\u003e\n\u003c/div\u003e\n\u003cdiv id=\"Sec7\" class=\"Section3\"\u003e\n\u003ch2\u003e4.1.2. Embedding Arabic text in a PIP-like background using color blending.\u003c/h2\u003e\n\u003cdiv class=\"BlockQuote\"\u003e\n\u003cp\u003eOnce the printed or handwritten Arabic text is generated (cf. Figures\u0026nbsp;4 and 5), it is embedded within a PIP\u0026ndash;style background via a three-phase color‐blending procedure: \u003cstrong\u003ecolor selection\u003c/strong\u003e, \u003cstrong\u003eword representation\u003c/strong\u003e, and \u003cstrong\u003eimage rendering\u003c/strong\u003e.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003ePhase 1: Color Selection\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eIn this phase, background and foreground hues are chosen from a predefined palette using a randomized\u0026mdash;but constrained\u0026mdash;selection process. Randomization is informed by empirical findings reported in [\u003cspan class=\"CitationRef\"\u003e39\u003c/span\u003e], which identified and excluded combinations with poor human performance (e.g., yellow-on‐white: 84% accuracy, 8.5s completion time; white‐on‐yellow: 79% accuracy, 9.8s) to ensure both usability and security aspects. Consequently, only color pairs demonstrating acceptable recognition rates and response times are retained.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003ePhase 2: Word Representation\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eRather than rendering text with standard fonts, each CAPTCHA word is depicted through an assembly of randomly sized and colored circles. These circles vary in chromaticity and luminance according to the selected color pair, producing unique character shapes in each instance. This stochastic, non-font‐based representation increases the difficulty for automated segmentation and character‐recognition algorithms, while remaining legible to human observers.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003ePhase 3: Image Rendering\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eIn the final phase, the colored-dot word representation is composited onto the chosen PIP-style background. The generator merges foreground and background layers, applies anti‐aliasing, and outputs the result as a high‐resolution, lossless image.\u003c/p\u003e\n\u003cp\u003eAccordingly, Fig.\u0026nbsp;6 provides the same examples used in Figs.\u0026nbsp;4 and 5 after embedding them into the PIP\u0026ndash;style background phase.\u003c/p\u003e\n\u003c/div\u003e\n\u003cdiv class=\"gridtable\"\u003e\u0026nbsp;\u003c/div\u003e\n\u003c/div\u003e\n\u003c/div\u003e\n\u003cdiv id=\"Sec8\" class=\"Section2\"\u003e\n\u003ch2\u003e4.2. Security Considerations\u003c/h2\u003e\n\u003cdiv class=\"BlockQuote\"\u003e\n\u003cp\u003eIn the proposed scheme, two key considerations must be addressed: resistance to current attacks and difficulty for color-blind bots. These aspects are discussed in the following subsections.\u003c/p\u003e\n\u003c/div\u003e\n\u003cdiv id=\"Sec9\" class=\"Section3\"\u003e\n\u003ch2\u003e4.2.1. Resistance to current attacks\u003c/h2\u003e\n\u003cdiv class=\"BlockQuote\"\u003e\n\u003cp\u003eThe proposed CAPTCHA system has been designed with an emphasis on resisting prevalent automated attacks, such as adversarial machine learning techniques [\u003cspan class=\"CitationRef\"\u003e2\u003c/span\u003e]. By leveraging the visual complexity inherent in PIP backgrounds\u0026mdash;combined with randomized text styles, positions, and color blending\u0026mdash;the system introduces a high degree of visual noise and distortion that hinders the effectiveness of these automated solvers. Furthermore, the integration of printed and handwritten text in varying visual configurations significantly increases the challenge for automated tools to accurately segment and interpret characters, thereby enhancing the CAPTCHA's robustness against sophisticated AI-driven attacks.\u003c/p\u003e\n\u003c/div\u003e\n\u003c/div\u003e\n\u003cdiv id=\"Sec10\" class=\"Section3\"\u003e\n\u003ch2\u003e4.2.2. Difficulty for color-blind bots\u003c/h2\u003e\n\u003cdiv class=\"BlockQuote\"\u003e\n\u003cp\u003eThe use of PIP principles in the CAPTCHA generation introduces an additional security layer specifically designed to exploit limitations in machine vision models that mimic human visual perception. In particular, PIPs are traditionally effective at distinguishing between individuals with normal color vision and those with color vision deficiencies. In this context, their application serves as a means to thwart bots that lack the nuanced color discrimination capabilities of human users. By embedding text in a way that is perceptible only through specific chromatic contrasts, the CAPTCHA ensures that bots\u0026mdash;which typically do not emulate color perception subtleties\u0026mdash;struggle to isolate and recognize the embedded text, further fortifying the system\u0026rsquo;s security.\u003c/p\u003e\n\u003c/div\u003e\n\u003c/div\u003e\n\u003c/div\u003e\n\u003cdiv id=\"Sec11\" class=\"Section2\"\u003e\n\u003ch2\u003e4.3. Usability Considerations\u003c/h2\u003e\n\u003cdiv class=\"BlockQuote\"\u003e\n\u003cp\u003eThe usability aspect of our proposed scheme should be evaluated with a focus on two main factors: human readability and accessibility for individuals with color vision variations. These aspects are explained in the following subsections.\u003c/p\u003e\n\u003c/div\u003e\n\u003cdiv id=\"Sec12\" class=\"Section3\"\u003e\n\u003ch2\u003e4.3.1. Human readability (tested on different user groups)\u003c/h2\u003e\n\u003cdiv class=\"BlockQuote\"\u003e\n\u003cp\u003eTo ensure the usability of the proposed CAPTCHA system, extensive testing was conducted across diverse user groups, including individuals of varying ages, educational backgrounds, and levels of digital literacy. This will be detailed in the following. The design prioritizes clear text visibility for human users by optimizing font styles, sizes, and color contrasts within the constraints of the PIP format. Despite the deliberate introduction of visual complexity to deter automated recognition, careful balancing was applied to maintain human readability. Feedback from pilot testing guided iterative refinements to improve clarity without compromising security, demonstrating that the CAPTCHA remains usable and user-friendly across a broad demographic.\u003c/p\u003e\n\u003c/div\u003e\n\u003c/div\u003e\n\u003cdiv id=\"Sec13\" class=\"Section3\"\u003e\n\u003ch2\u003e4.3.2. Color vision variations and accessibility\u003c/h2\u003e\n\u003cdiv class=\"BlockQuote\"\u003e\n\u003cp\u003eThe system also accounts for variations in human color perception, particularly those related to common forms of color vision deficiency. While PIP principles inherently exploit color differentiation to enhance security, design parameters were adjusted to preserve a baseline level of accessibility. Specific color combinations known to be problematic for users with red-green or blue-yellow deficiencies were avoided, in alignment with empirical findings from color vision research. Additionally, the CAPTCHA's difficulty was calibrated to ensure that users with mild to moderate color vision impairments could still successfully complete the challenge, promoting inclusivity without undermining the CAPTCHA\u0026rsquo;s resistance to bots.\u003c/p\u003e\n\u003c/div\u003e\n\u003c/div\u003e\n\u003c/div\u003e"},{"header":"5. Experimental evaluation","content":"\u003cdiv class=\"BlockQuote\"\u003e\n\u003cp\u003eThis section presents an evaluation of the proposed approach from both security and usability perspectives, based on a structured experimental study. The experimental setup and procedures for each aspect are described in detail.\u003c/p\u003e\n\u003c/div\u003e\n\u003cdiv id=\"Sec15\" class=\"Section2\"\u003e\n\u003ch2\u003e5.1. Experiment setup\u003c/h2\u003e\n\u003cp\u003eTo assess the usability of the proposed Arabic CAPTCHA generation method, a within-subject laboratory study was conducted. The experiment was carefully designed to evaluate user interaction, readability, and accessibility across different CAPTCHA formats. Participant recruitment began on 15 April 2025 and ended on 30 May 2025. The following subsections detail the key components of the experimental setup. It is important noting that all methods were carried out in accordance with relevant guidelines and regulations. Also, all experimental protocols were approved by a named institutional. Besides, informed consent was obtained from all subjects and/or their legal guardian.\u003c/p\u003e\n\u003cdiv id=\"Sec16\" class=\"Section3\"\u003e\n\u003ch2\u003e5.1.1. Design\u003c/h2\u003e\n\u003cdiv class=\"BlockQuote\"\u003e\n\u003cp\u003eA within-subject design was employed to ensure that each participant experienced all CAPTCHA variants under controlled conditions, allowing for direct comparisons of usability factors.\u003c/p\u003e\n\u003c/div\u003e\n\u003c/div\u003e\n\u003cdiv id=\"Sec17\" class=\"Section3\"\u003e\n\u003ch2\u003e5.1.2. Description of dataset\u003c/h2\u003e\n\u003cdiv class=\"BlockQuote\"\u003e\n\u003cp\u003eThe dataset comprised both meaningful and meaningless Arabic text samples in printed and handwritten forms. Printed samples were generated using the method described in [\u003cspan class=\"CitationRef\"\u003e21\u003c/span\u003e], while meaningless handwritten samples were produced using the tool outlined in [\u003cspan class=\"CitationRef\"\u003e38\u003c/span\u003e]. The generation of meaningful handwritten words followed the approach introduced in [\u003cspan class=\"CitationRef\"\u003e3\u003c/span\u003e]. In total, 60 samples were generated to support the evaluation process.\u003c/p\u003e\n\u003c/div\u003e\n\u003c/div\u003e\n\u003cdiv id=\"Sec18\" class=\"Section3\"\u003e\n\u003ch2\u003e5.1.3. PIP design parameters:\u003c/h2\u003e\n\u003cdiv class=\"BlockQuote\"\u003e\n\u003cp\u003eThe design of the pseudoisochromatic CAPTCHA plates was carefully configured to balance security and human readability. Several visual parameters were adjusted, each playing a critical role in shaping the user experience and resistance to automated attacks. These parameters are outlined in the following points.\u003c/p\u003e\n\u003c/div\u003e\n\u003cp\u003e\u003cstrong\u003eColor Schemes\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eA range of foreground and background color combinations was employed to ensure sufficient contrast for users with normal vision while remaining challenging for automated solvers. Color pairings known to be problematic for readability (e.g., white-on-yellow or yellow-on-white, as identified in [\u003cspan class=\"CitationRef\"\u003e39\u003c/span\u003e]) were explicitly avoided. Instead, empirically validated combinations were used to maintain visual distinction and legibility.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eNoise Density\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eBackground noise elements\u0026mdash;random circles varying in chromaticity, luminance, and size\u0026mdash;were introduced to obscure the CAPTCHA text from machine-based recognition systems. The noise level was carefully calibrated to avoid excessive interference with human readability, particularly for users with mild visual impairments.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eDistractor Types\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eAdditional distractor elements, including overlapping shapes and color gradients, were incorporated to further complicate segmentation and character recognition by bots. Figure\u0026nbsp;\u003cspan class=\"InternalRef\"\u003e7\u003c/span\u003e illustrates a sample of implementing this parameter. These distractors were randomized per image to reduce pattern predictability and increase variability across samples.\u003c/p\u003e\n\u003cdiv class=\"BlockQuote\"\u003e\n\u003cp\u003eCollectively, these parameters were optimized to create visually complex CAPTCHA images that remain accessible to human users while significantly hindering machine recognition capabilities.\u003c/p\u003e\n\u003c/div\u003e\n\u003cp\u003e\u0026nbsp;\u003c/p\u003e\n\u003c/div\u003e\n\u003cdiv id=\"Sec19\" class=\"Section3\"\u003e\n\u003ch2\u003e5.1.4. Participant demographics\u003c/h2\u003e\n\u003cdiv class=\"BlockQuote\"\u003e\n\u003cp\u003eTo capture relevant background information, participants were asked to complete a brief demographic survey prior to beginning the experiment. The survey included the following items: prior experience with Arabic CAPTCHAs (i.e., whether the participant had previously encountered or used Arabic CAPTCHAs), gender, age group, and education level. This information was collected to analyze potential variations in usability performance and perception based on user characteristics, and to ensure a diverse and representative sample of Arabic-speaking users.\u003c/p\u003e\n\u003c/div\u003e\n\u003c/div\u003e\n\u003cdiv id=\"Sec20\" class=\"Section3\"\u003e\n\u003ch2\u003e5.1.5. Baseline CAPTCHA methods for comparison\u003c/h2\u003e\n\u003cdiv class=\"BlockQuote\"\u003e\n\u003cp\u003eTo evaluate the impact of the PIP embedding, the original generated Arabic text samples\u0026mdash;both printed and handwritten\u0026mdash;were used as baseline CAPTCHA methods, as shown previously in Figs.\u0026nbsp;\u003cspan class=\"InternalRef\"\u003e3\u003c/span\u003e and 4. These unembedded samples represent standard, high-contrast text images without any background noise or color blending. By comparing user responses and bot resistance between these plain text images and their PIP-enhanced counterparts, we were able to quantify the trade-offs in usability and security introduced by our proposed method. Specifically, the main reason behind using the original generated samples is as follows.\u003c/p\u003e\n\u003c/div\u003e\n\u003cp\u003e\u003cstrong\u003eControl Variable\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eThe original samples (plain printed or handwritten text) serve as a controlled version of the CAPTCHA, isolating the effect of the PIP-based embedding.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eUsability Benchmark\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eComparing user performance (e.g., recognition accuracy and time) on plain samples versus PIP-enhanced samples helps quantify any usability trade-offs.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eSecurity Contrast\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eBots may find the original text easier to segment and recognize, so the baseline also allows a direct assessment of how much the PIP-based method enhances resistance to automated attacks.\u003c/p\u003e\n\u003c/div\u003e\n\u003cdiv id=\"Sec21\" class=\"Section3\"\u003e\n\u003ch2\u003e5.1.6. System\u003c/h2\u003e\n\u003cdiv class=\"BlockQuote\"\u003e\n\u003cp\u003eAn online testing interface was developed using Weavely Forms and the Figma platform, providing participants with interactive access to the proposed CAPTCHA samples and recording their performance metrics and feedback.\u003c/p\u003e\n\u003c/div\u003e\n\u003c/div\u003e\n\u003c/div\u003e\n\u003cdiv id=\"Sec22\" class=\"Section2\"\u003e\n\u003ch2\u003e5.2. Experiment procedure\u003c/h2\u003e\n\u003cdiv class=\"BlockQuote\"\u003e\n\u003cp\u003eThis section outlines the detailed methodology of the experiment, including participant instructions, procedures for evaluating both security and usability, and the data collected.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003ei. Instructions to participants\u003c/strong\u003e:\u003c/p\u003e\n\u003c/div\u003e\n\u003cdiv class=\"BlockQuote\"\u003e\n\u003cp\u003eAt the outset, participants completed a demographic survey capturing relevant background information. They were then informed that they would be presented with 60 Arabic text-based CAPTCHA samples displayed in the following order: (1) printed-based meaningful and meaningless texts, then (2) handwritten-based meaningful and meaningless texts, alongside baseline samples representing each one of them. Upon navigating to the CAPTCHA challenge page, participants were instructed to carefully observe each displayed CAPTCHA and enter the recognized characters into the provided input field, submitting their response by clicking the submit button. It is important to note that no feedback regarding answer correctness or hints was provided during the task. Participants were also advised to minimize distractions and maintain focus throughout the experiment.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eii. Security experiment procedure\u003c/strong\u003e\u003c/p\u003e\n\u003c/div\u003e\n\u003cdiv class=\"BlockQuote\"\u003e\n\u003cp\u003eTo assess the security robustness of the proposed CAPTCHA system, we utilized the Google Vision API [\u003cspan class=\"CitationRef\"\u003e40\u003c/span\u003e] as a benchmark for evaluating resistance against recognition and segmentation attacks. Given its widespread use and high accuracy, the Cloud Vision API served as an appropriate tool for this purpose. All 60 CAPTCHA samples were processed through the Google Vision engine, and the resulting text labels were categorized into four recognition outcomes: completely recognized, partially recognized, incorrectly recognized, and not recognized. Table\u0026nbsp;\u003cspan class=\"InternalRef\"\u003e2\u003c/span\u003e summarizes these categories along with their descriptions.\u003c/p\u003e\n\u003c/div\u003e\n\u003cdiv class=\"gridtable\"\u003e\n\u003ctable id=\"Tab2\" border=\"1\"\u003e\u003ccaption\u003e\n\u003cdiv class=\"CaptionNumber\"\u003eTable 2\u003c/div\u003e\n\u003cdiv class=\"CaptionContent\"\u003e\n\u003cp\u003eRecognition Outputs Categories\u003c/p\u003e\n\u003c/div\u003e\n\u003c/caption\u003e\n\u003cthead\u003e\n\u003ctr\u003e\n\u003cth align=\"left\"\u003e\n\u003cp\u003eRecognition Category (RC)\u003c/p\u003e\n\u003c/th\u003e\n\u003cth align=\"left\"\u003e\n\u003cp\u003eDescription\u003c/p\u003e\n\u003c/th\u003e\n\u003c/tr\u003e\n\u003c/thead\u003e\n\u003ctbody\u003e\n\u003ctr\u003e\n\u003ctd align=\"left\"\u003e\n\u003cp\u003eCompletely (C)\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd align=\"left\"\u003e\n\u003cp\u003eAll characters correctly recognized\u003c/p\u003e\n\u003c/td\u003e\n\u003c/tr\u003e\n\u003ctr\u003e\n\u003ctd align=\"left\"\u003e\n\u003cp\u003ePartially (P)\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd align=\"left\"\u003e\n\u003cp\u003eSome characters correctly recognized\u003c/p\u003e\n\u003c/td\u003e\n\u003c/tr\u003e\n\u003ctr\u003e\n\u003ctd align=\"left\"\u003e\n\u003cp\u003eIncorrectly (I)\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd align=\"left\"\u003e\n\u003cp\u003eAll characters incorrectly recognized\u003c/p\u003e\n\u003c/td\u003e\n\u003c/tr\u003e\n\u003ctr\u003e\n\u003ctd align=\"left\"\u003e\n\u003cp\u003eNot recognized (NR)\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd align=\"left\"\u003e\n\u003cp\u003eNo characters recognized\u003c/p\u003e\n\u003c/td\u003e\n\u003c/tr\u003e\n\u003c/tbody\u003e\n\u003c/table\u003e\n\u003c/div\u003e\n\u003c/div\u003e\n\u003cdiv class=\"Section2\"\u003e\u0026nbsp;\u003c/div\u003e\n\u003cdiv class=\"Section2\"\u003e\u003cstrong\u003eiii. Usability experiment procedure\u003c/strong\u003e:\u003cbr /\u003e\n\u003cdiv class=\"BlockQuote\"\u003e\n\u003cp\u003eThe usability of the proposed CAPTCHA system was evaluated through both quantitative and qualitative metrics. Quantitative evaluation involved measuring the efficiency and effectiveness of the system. Efficiency was quantified by recording the time (in seconds) taken by participants to accurately type the CAPTCHA text\u0026mdash;from clicking the \u0026lsquo;Start\u0026rsquo; button to submitting their response. Effectiveness was measured as the success rate, defined by the proportion of CAPTCHAs correctly solved by the participants. For the qualitative metric, data were obtained via a post-experiment survey assessing user satisfaction and perceived difficulty.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eiv. Collected data\u003c/strong\u003e:\u003c/p\u003e\n\u003c/div\u003e\n\u003cdiv class=\"BlockQuote\"\u003e\n\u003cp\u003eDuring the experiment, key performance indicators including success rates, response times, and user satisfaction scores were systematically recorded and analyzed to provide a comprehensive assessment of the proposed approach.\u003c/p\u003e\n\u003c/div\u003e\n\u003c/div\u003e"},{"header":"6. Results","content":"\u003cdiv class=\"BlockQuote\"\u003e\n\u003cp\u003eIn the experimental study, all participants completed the tasks successfully. The security evaluation, which shows the robustness level against such attacks, of proposed approach is presented. This is followed by the usability results which includes the efficiency, effectiveness, and satisfaction rates of the participants.\u003c/p\u003e\n\u003c/div\u003e\n\u003cdiv id=\"Sec24\" class=\"Section2\"\u003e\n\u003ch2\u003e6.1. Security evaluation results\u003c/h2\u003e\n\u003cdiv class=\"BlockQuote\"\u003e\n\u003cp\u003eTable\u0026nbsp;\u003cspan class=\"InternalRef\"\u003e3\u003c/span\u003e summarizes the total percentage of robustness level of the proposed approach against such sophisticated attacks using Google Vision API [\u003cspan class=\"CitationRef\"\u003e40\u003c/span\u003e]. In particular, the results showed the robustness level of printed-based for meaningful, meaningless and baseline text. Also the robustness level of handwritten-based for meaningful, meaningless and baseline texts are shown.\u003c/p\u003e\n\u003c/div\u003e\n\u003cdiv class=\"gridtable\"\u003e\n\u003cdiv class=\"colspec\" align=\"left\"\u003e\u0026nbsp;\u003c/div\u003e\n\u003cdiv class=\"colspec\" align=\"left\"\u003e\u0026nbsp;\u003c/div\u003e\n\u003ctable id=\"Tab3\" border=\"1\"\u003e\u003ccaption\u003e\n\u003cdiv class=\"CaptionNumber\"\u003eTable 3\u003c/div\u003e\n\u003cdiv class=\"CaptionContent\"\u003e\n\u003cp\u003eThe results of security evaluation\u003c/p\u003e\n\u003c/div\u003e\n\u003c/caption\u003e\n\u003cthead\u003e\n\u003ctr\u003e\n\u003cth rowspan=\"3\" align=\"left\"\u003e\n\u003cp\u003eRC\u003c/p\u003e\n\u003c/th\u003e\n\u003cth colspan=\"4\" align=\"left\"\u003e\n\u003cp\u003ePrinted-based text\u003c/p\u003e\n\u003c/th\u003e\n\u003cth colspan=\"4\" align=\"left\"\u003e\n\u003cp\u003eHandwritten-based text\u003c/p\u003e\n\u003c/th\u003e\n\u003c/tr\u003e\n\u003ctr\u003e\n\u003cth rowspan=\"2\" align=\"left\"\u003e\n\u003cp\u003eMeaningful\u003c/p\u003e\n\u003c/th\u003e\n\u003cth rowspan=\"2\" align=\"left\"\u003e\n\u003cp\u003eMeaningless\u003c/p\u003e\n\u003c/th\u003e\n\u003cth colspan=\"2\" align=\"left\"\u003e\n\u003cp\u003eBaseline\u003c/p\u003e\n\u003c/th\u003e\n\u003cth rowspan=\"2\" align=\"left\"\u003e\n\u003cp\u003eMeaningful\u003c/p\u003e\n\u003c/th\u003e\n\u003cth rowspan=\"2\" align=\"left\"\u003e\n\u003cp\u003eMeaningless\u003c/p\u003e\n\u003c/th\u003e\n\u003cth colspan=\"2\" align=\"left\"\u003e\n\u003cp\u003eBaseline\u003c/p\u003e\n\u003c/th\u003e\n\u003c/tr\u003e\n\u003c/thead\u003e\n\u003ctbody\u003e\n\u003ctr\u003e\n\u003ctd align=\"left\"\u003e\n\u003cp\u003eMeaningful\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd align=\"left\"\u003e\n\u003cp\u003eMeaningless\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd align=\"left\"\u003e\n\u003cp\u003eMeaningful\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd align=\"left\"\u003e\n\u003cp\u003eMeaningless\u003c/p\u003e\n\u003c/td\u003e\n\u003c/tr\u003e\n\u003ctr\u003e\n\u003ctd align=\"left\"\u003e\n\u003cp\u003eC\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd align=\"left\"\u003e\n\u003cp\u003e0%\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd align=\"left\"\u003e\n\u003cp\u003e0%\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd align=\"left\"\u003e\n\u003cp\u003e90%\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd align=\"left\"\u003e\n\u003cp\u003e100%\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd align=\"left\"\u003e\n\u003cp\u003e0\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd align=\"left\"\u003e\n\u003cp\u003e0\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd align=\"left\"\u003e\n\u003cp\u003e70%\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd align=\"left\"\u003e\n\u003cp\u003e60%\u003c/p\u003e\n\u003c/td\u003e\n\u003c/tr\u003e\n\u003ctr\u003e\n\u003ctd align=\"left\"\u003e\n\u003cp\u003eP\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd align=\"left\"\u003e\n\u003cp\u003e0%\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd align=\"left\"\u003e\n\u003cp\u003e0%\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd align=\"left\"\u003e\n\u003cp\u003e10%\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd align=\"left\"\u003e\n\u003cp\u003e0%\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd align=\"left\"\u003e\n\u003cp\u003e0\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd align=\"left\"\u003e\n\u003cp\u003e0\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd align=\"left\"\u003e\n\u003cp\u003e20%\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd align=\"left\"\u003e\n\u003cp\u003e40%\u003c/p\u003e\n\u003c/td\u003e\n\u003c/tr\u003e\n\u003ctr\u003e\n\u003ctd align=\"left\"\u003e\n\u003cp\u003eI\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd align=\"left\"\u003e\n\u003cp\u003e0%\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd align=\"left\"\u003e\n\u003cp\u003e0%\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd align=\"left\"\u003e\n\u003cp\u003e0%\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd align=\"left\"\u003e\n\u003cp\u003e0%\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd align=\"left\"\u003e\n\u003cp\u003e0\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd align=\"left\"\u003e\n\u003cp\u003e0\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd align=\"left\"\u003e\n\u003cp\u003e0%\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd align=\"left\"\u003e\n\u003cp\u003e0%\u003c/p\u003e\n\u003c/td\u003e\n\u003c/tr\u003e\n\u003ctr\u003e\n\u003ctd align=\"left\"\u003e\n\u003cp\u003eN\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd align=\"left\"\u003e\n\u003cp\u003e100%\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd align=\"left\"\u003e\n\u003cp\u003e100%\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd align=\"left\"\u003e\n\u003cp\u003e0%\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd align=\"left\"\u003e\n\u003cp\u003e0%\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd align=\"left\"\u003e\n\u003cp\u003e100%\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd align=\"left\"\u003e\n\u003cp\u003e100%\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd align=\"left\"\u003e\n\u003cp\u003e10%\u003c/p\u003e\n\u003c/td\u003e\n\u003ctd align=\"left\"\u003e\n\u003cp\u003e0%\u003c/p\u003e\n\u003c/td\u003e\n\u003c/tr\u003e\n\u003c/tbody\u003e\n\u003c/table\u003e\n\u003c/div\u003e\n\u003c/div\u003e\n\u003cdiv id=\"Sec25\" class=\"Section2\"\u003e\n\u003ch2\u003e6.2. Usability evaluation results\u003c/h2\u003e\n\u003cdiv class=\"BlockQuote\"\u003e\n\u003cp\u003eThe usability experiment is performed after the security experiment. This section shows the results of testing the efficiency, effectiveness, and the satisfaction with the proposed approach.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003e6.2.1. Efficiency\u003c/strong\u003e: Fig.\u0026nbsp;\u003cspan class=\"InternalRef\"\u003e8\u003c/span\u003e illustrates the average duration (in seconds) participants required to correctly input the CAPTCHA text. The bars represent the average time taken (in seconds) to solve each CAPTCHA type. The dashed lines show the overall average efficiency: 7.75 seconds for the Baseline scheme and 23.75 seconds for the Proposed scheme. The increased time reflects the enhanced robustness and complexity introduced in the Proposed CAPTCHA design.\u003c/p\u003e\n\u003c/div\u003e\n\u003cp\u003e\u003cstrong\u003e6.2.2 Effectiveness:\u003c/strong\u003e The success rate of correctly solving of released CAPTCHA samples is shown in Fig.\u0026nbsp;\u003cspan class=\"InternalRef\"\u003e9\u003c/span\u003e. The vertical bars represent the average success rates for each condition, while the dashed lines indicate the overall average effectiveness for the Baseline (99.0%) and Proposed (90.8%) schemes. The results highlight a slight decrease in effectiveness for the Proposed scheme due to increased robustness, yet maintain high human solvability across all text types.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003e6.2.3.\u0026nbsp;Satisfaction\u003c/strong\u003e: Upon completing all assigned tasks, participants were asked to respond to a set of questions assessing their experience with the proposed CAPTCHA approach. The survey questions were grouped into four key categories: \u003cem\u003ePerceived Readability and Clarity\u003c/em\u003e, \u003cem\u003eEase of Solving and Interaction\u003c/em\u003e, \u003cem\u003eUser Frustration or Fatigue\u003c/em\u003e, and \u003cem\u003eAesthetic and Visual Impression\u003c/em\u003e.\u003c/p\u003e\n\u003cdiv class=\"BlockQuote\"\u003e\n\u003cp\u003eIn the first category, participants were asked questions such as: \u003cem\u003e\"It was easy to read the characters in the CAPTCHA images\"\u003c/em\u003e and \u003cem\u003e\"The contrast between the text and background was sufficient for me to read the CAPTCHA.\"\u003c/em\u003e For \u003cem\u003eEase of Solving and Interaction\u003c/em\u003e, the question was: \u003cem\u003e\"I was able to recognize and type the CAPTCHA text without much effort.\"\u003c/em\u003e The third category, \u003cem\u003eUser Frustration or Fatigue\u003c/em\u003e, included: \u003cem\u003e\"Solving the CAPTCHAs made me feel frustrated or fatigued.\"\u003c/em\u003e Finally, under \u003cem\u003eAesthetic and Visual Impression\u003c/em\u003e, the statement was: \u003cem\u003e\"The design of the CAPTCHA was visually appealing.\"\u003c/em\u003e\u003c/p\u003e\n\u003cp\u003eResponses were rated on a five-point Likert scale ranging from 1 (\u003cem\u003eStrongly Disagree\u003c/em\u003e) to 5 (\u003cem\u003eStrongly Agree\u003c/em\u003e), represented by a star rating system. A summary of the survey results is presented in Fig.\u0026nbsp;10.\u003c/p\u003e\n\u003c/div\u003e\n\u003c/div\u003e"},{"header":"7. Discussion","content":"\u003cp\u003e\u003cdiv class=\"BlockQuote\"\u003e\u003cp\u003eThis section synthesizes the findings of the experimental evaluation and reflects on the broader implications of the proposed PIP-style Arabic CAPTCHA approach. By analyzing both security and usability outcomes, we aim to assess the effectiveness of the method and identify its practical strengths and limitations. Furthermore, we discuss specific considerations relevant to Arabic-language contexts and propose directions for future enhancement. The following subsections explore key trade-offs, advantages, constraints, and opportunities for improvement.\u003c/p\u003e\u003c/div\u003e\u003c/p\u003e\u003cdiv id=\"Sec27\" class=\"Section2\"\u003e\u003ch2\u003e7.1. Trade-offs between security and usability\u003c/h2\u003e\u003cp\u003e\u003cdiv class=\"BlockQuote\"\u003e\u003cp\u003eThe findings of this study highlight an inherent trade-off between CAPTCHA security and usability. The integration of PIP-style backgrounds significantly enhances resistance to automated recognition systems, particularly by impeding character segmentation and distorting character shapes. However, this enhancement comes at the cost of reduced usability for human users. For instance, handwritten text samples embedded in PIP backgrounds demonstrated robust security performance but required longer response times and exhibited slightly lower success rates among participants, as illustrated in Fig.\u0026nbsp;\u003cspan refid=\"Fig3\" class=\"InternalRef\"\u003e8\u003c/span\u003e. In contrast, both meaningful and meaningless printed texts offered a more user-friendly experience while still maintaining substantial resistance to segmentation and recognition attacks. Given that printed text samples outperformed handwritten samples in terms of both efficiency and effectiveness, their use in the proposed scheme is recommended. Nonetheless, further research is needed to validate these findings and explore potential optimizations. Overall, achieving an optimal balance between security and usability remains a central challenge in CAPTCHA design, particularly for languages with complex scripts such as Arabic.\u003c/p\u003e\u003c/div\u003e\u003c/p\u003e\u003c/div\u003e\u003cdiv id=\"Sec28\" class=\"Section2\"\u003e\u003ch2\u003e7.2. Strengths of the PIP approach in Arabic contexts\u003c/h2\u003e\u003cp\u003e\u003cdiv class=\"BlockQuote\"\u003e\u003cp\u003eThe proposed approach capitalizes on the visual and linguistic intricacies of the Arabic script, which already poses challenges to OCR systems. When coupled with the PIP framework, these inherent complexities are further amplified, leading to stronger protection against bot-based attacks, as illustrated in Table\u0026nbsp;\u003cspan refid=\"Tab3\" class=\"InternalRef\"\u003e3\u003c/span\u003e. Moreover, the stylistic variations introduced through randomized fonts, character sizes, and chromatic noise patterns enhance both unpredictability and visual entropy. In user testing, Arabic-speaking participants expressed a positive response toward the novel presentation style, noting its cultural and linguistic relevance, as summarized in Fig.\u0026nbsp;10. This underscores the adaptability of the PIP mechanism to Arabic-language applications, where traditional Latin-based CAPTCHA systems often fall short in both usability and effectiveness.\u003c/p\u003e\u003c/div\u003e\u003c/p\u003e\u003c/div\u003e\u003cdiv id=\"Sec29\" class=\"Section2\"\u003e\u003ch2\u003e7.3. Limitations: potential issues for color-blind users, complex generation process\u003c/h2\u003e\u003cp\u003e\u003cdiv class=\"BlockQuote\"\u003e\u003cp\u003eDespite its strengths, the PIP-based CAPTCHA system is not without limitations. One critical concern is accessibility for users with color vision deficiencies. Although the design avoids color combinations known to be problematic, the inherently chromatic nature of PIP backgrounds can still impede recognition for color-blind users. Future iterations should include an accessibility mode or offer contrast-enhancing alternatives. Additionally, the generation process is relatively complex, involving fine-tuned color selection, shape rendering, and integration with character layers. This complexity may pose implementation challenges for web developers or system integrators lacking advanced image processing capabilities.\u003c/p\u003e\u003c/div\u003e\u003c/p\u003e\u003c/div\u003e\u003cdiv id=\"Sec30\" class=\"Section2\"\u003e\u003ch2\u003e7.4. Future enhancements\u003c/h2\u003e\u003cp\u003e\u003cdiv class=\"BlockQuote\"\u003e\u003cp\u003eTo address these limitations and extend the system\u0026rsquo;s applicability, several enhancements are proposed. First, the development of \u003cb\u003edynamic plate generators\u003c/b\u003e\u0026mdash;capable of adjusting color schemes and visual density based on real-time user performance\u0026mdash;could improve accessibility and personalization. Second, \u003cb\u003eadaptive contrast techniques\u003c/b\u003e could be integrated to detect and compensate for potential vision deficiencies, ensuring that CAPTCHA elements remain distinguishable across a wider user base. Lastly, further research into \u003cb\u003emultimodal CAPTCHA systems\u003c/b\u003e, combining PIP visuals with auditory or haptic cues, may provide inclusive alternatives while preserving high security standards.\u003c/p\u003e\u003c/div\u003e\u003c/p\u003e\u003c/div\u003e"},{"header":"8. Conclusion","content":"\u003cp\u003e\u003cdiv class=\"BlockQuote\"\u003e\u003cp\u003eThis study introduced a novel Arabic CAPTCHA generation method that incorporates Pseudoisochromatic Plate (PIP) principles to improve resistance against automated recognition attacks while maintaining human readability. The proposed framework demonstrated strong robustness, particularly through the integration of visual complexity via color blending and randomized character representation. Experimental results confirmed the system\u0026rsquo;s effectiveness in countering machine-based attacks, including those leveraging advanced recognition tools like the Google Vision API. Furthermore, usability evaluations revealed that while the handwritten PIP-embedded CAPTCHAs offer heightened security, printed text variants provide superior usability in terms of efficiency and effectiveness. This trade-off emphasizes the ongoing challenge of balancing security and accessibility in CAPTCHA design\u0026mdash;especially for complex scripts such as Arabic. The proposed system offers a significant contribution to the field by introducing perceptual-based obfuscation mechanisms that are linguistically and culturally adapted. Future research will focus on enhancing accessibility for users with color vision deficiencies and refining dynamic generation techniques to support broader inclusivity and implementation in real-world applications.\u003c/p\u003e\u003c/div\u003e\u003c/p\u003e"},{"header":"Declarations","content":"\u003ch2\u003eFunding:\u003c/h2\u003e\u003cp\u003eThe Researcher would like to thank the Deanship of Graduate Studies and Scientific Research at Qassim University for financial support (QU-APC-2025).\u003c/p\u003e\u003ch2\u003eAuthor Contribution\u003c/h2\u003e\u003cp\u003eConceptualization, S. A. Alsuhibany.; methodology, S. A. Alsuhibany.; software, S. A. Alsuhibany.; validation, S. A. Alsuhibany.; formal analysis, S. A. Alsuhibany.; investigation, S. A. Alsuhibany.; resources, S. A. Alsuhibany.; data curation, S. A. Alsuhibany.; writing\u0026mdash;original draft preparation, S. A. Alsuhibany.; writing\u0026mdash;review and editing, S. A. Alsuhibany.; visualization, S. A. Alsuhibany\u003c/p\u003e\u003ch2\u003eAcknowledgement\u003c/h2\u003e\u003cp\u003eThe Researcher would like to thank the Deanship of Graduate Studies and Scientific Research at Qassim University for financial support (QU-APC-2025)\u003c/p\u003e\u003ch2\u003eData Availability\u003c/h2\u003e\u003cp\u003eThe data that support the findings of this study are available from the corresponding author, Suliman A. Alsuhibany, upon reasonable request.\u003c/p\u003e"},{"header":"References","content":"\u003col\u003e\u003cli\u003e\u003cspan\u003eAlsuhibany, S. A. \u0026amp; Parvez, M. T. Secure Arabic Handwritten CAPTCHA Generation Using OCR Operations. In: 2016 15th International Conference on Frontiers in Handwriting Recognition (ICFHR), pp. 126\u0026ndash;131. (2016). \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.1109/ICFHR.2016.31\u003c/span\u003e\u003cspan address=\"10.1109/ICFHR.2016.31\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eAlsuhibany, S. A. A survey on adversarial perturbations and attacks on CAPTCHAs. \u003cem\u003eAppl. Sci.\u003c/em\u003e \u003cb\u003e13\u003c/b\u003e (7), 4602. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.3390/app13074602\u003c/span\u003e\u003cspan address=\"10.3390/app13074602\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e (2023).\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eAlrasheed, G. \u0026amp; Alsuhibany, S. A. Enhancing Security of Online Interfaces: Adversarial Handwritten Arabic CAPTCHA Generation. \u003cem\u003eAppl. Sci.\u003c/em\u003e \u003cb\u003e15\u003c/b\u003e (6), 2972. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.3390/app15062972\u003c/span\u003e\u003cspan address=\"10.3390/app15062972\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e (2025).\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eShirali-Shahreza, M. H. \u0026amp; Shirali-Shahreza, M. \u003cem\u003ePersian/Arabic Baffletext CAPTCHA J. UCS\u003c/em\u003e \u003cb\u003e12\u003c/b\u003e(12), 1783\u0026ndash;1796. (2006).\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eShahreza, M. S. Verifying Spam SMS by Arabic CAPTCHA. In: 2nd Int. Conf. on Information \u0026amp; Communication Technologies (ICT), vol. 1, pp. 78\u0026ndash;83. (2006). \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.1109/ICT.2006.253141\u003c/span\u003e\u003cspan address=\"10.1109/ICT.2006.253141\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eKhan, B., Alghathbar, K. S., Khan, M. K., AlKelabi, A. M. \u0026amp; AlAjaji, A. Using Arabic CAPTCHA for cyber security. In: Security Technology, Disaster Recovery and Business Continuity, pp. 8\u0026ndash;17. Springer. (2010). \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.1007/978-3-642-16644-0_2\u003c/span\u003e\u003cspan address=\"10.1007/978-3-642-16644-0_2\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eBakry, M., Khamis, M., Abdennadher, S. \u0026amp; Alhodathi, H. Arecaptcha: Outsourcing Arabic text digitization to native speakers. In: 11th IAPR Int. Workshop on Document Analysis Systems (DAS), pp. 304\u0026ndash;308. (2014). \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.1109/DAS.2014.66\u003c/span\u003e\u003cspan address=\"10.1109/DAS.2014.66\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eAbubaker, H., Salah, K., Al-Muhairi, H. \u0026amp; Bentiba, A. Cloud-based Arabic reCAPTCHA service: Design and architecture. In: AICCSA, pp. 1\u0026ndash;6. (2015). \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.1109/AICCSA.2015.7507200\u003c/span\u003e\u003cspan address=\"10.1109/AICCSA.2015.7507200\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eAbubaker, H., Salah, K., Al-Muhairi, H. \u0026amp; Bentiba, A. Architectural design of a cloud-based reCAPTCHA service. In: 12th Int. Conf. on Innovations in Information Technology (IIT), pp. 1\u0026ndash;6. (2016). \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.1109/INNOVATIONS.2016.7880041\u003c/span\u003e\u003cspan address=\"10.1109/INNOVATIONS.2016.7880041\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eRodriguez-Carmona, M. Pseudoisochromatic Plates. In: (ed Luo, R.) Encyclopedia of Color Science and Technology, Springer, Berlin, Heidelberg, 1\u0026ndash;2. (2014).\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eParvez, M. T. \u0026amp; Alsuhibany, S. A. Challenges and opportunities for Arabic CAPTCHAs. \u003cem\u003eMultimed Tools Appl.\u003c/em\u003e \u003cb\u003e83\u003c/b\u003e (5), 14047\u0026ndash;14062. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.1007/s11042-023-15847-9\u003c/span\u003e\u003cspan address=\"10.1007/s11042-023-15847-9\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e (2024).\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eAlsuhibany, S. A. \u0026amp; Parvez, M. T. Attack-filtered interactive Arabic CAPTCHAs. \u003cem\u003eJ. Inf. Secur. Appl.\u003c/em\u003e \u003cb\u003e70\u003c/b\u003e, 103318. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.1016/j.jisa.2022.103318\u003c/span\u003e\u003cspan address=\"10.1016/j.jisa.2022.103318\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e (2022).\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eKhan, B., Alghathbar, K., Khan, M. K., AlKelabi, A. M. \u0026amp; Alajaji, A. Cyber security using Arabic CAPTCHA scheme. \u003cem\u003eInt. Arab. J. Inf. Technol.\u003c/em\u003e \u003cb\u003e10\u003c/b\u003e (1), 76\u0026ndash;84 (2013).\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eBanday, M. T. \u0026amp; Sheikh, S. A. \u003cem\u003eCAPTCHA Chall. Int. J. Web Portals\u003c/em\u003e \u003cb\u003e7\u003c/b\u003e(1), 1\u0026ndash;27. (2015).\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eSulaiman, M. T. \u0026amp; Hassan, N. F. Generate Random Arabic Characters and Numbers for CAPTCHA. \u003cem\u003eIraqi J. Sci.\u003c/em\u003e \u003cb\u003e58\u003c/b\u003e (4B), 2427\u0026ndash;2437 (2017).\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eAlsuhibany, S. A. \u0026amp; Alhodathi, H. A. A Parametric Study of Arabic Text-Based CAPTCHA Difficulty for Humans. \u003cem\u003eIntell. Autom. Soft Comput.\u003c/em\u003e \u003cb\u003e31\u003c/b\u003e (1), 523\u0026ndash;537. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.32604/iasc.2022.017266\u003c/span\u003e\u003cspan address=\"10.32604/iasc.2022.017266\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e (2022).\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eSulaiman, M. T. Propose an Arabic CAPTCHA System based on Chaotic Maps. \u003cem\u003eEng. Technol. J.\u003c/em\u003e \u003cb\u003e36\u003c/b\u003e (1B), 13\u0026ndash;21 (2018).\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eAlsuhibany, S. A., Parvez, M. T., Alrobah, N., Almohaimeed, F. \u0026amp; Alduayji, S. Evaluating robustness of Arabic CAPTCHAs. In: 2nd Int. Conf. on Anti-Cyber Crimes (ICACC), pp. 81\u0026ndash;86. (2017). \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.1109/ICACC.2017.7914984\u003c/span\u003e\u003cspan address=\"10.1109/ICACC.2017.7914984\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eShirali-Shahreza, M. H. \u0026amp; Shirali-Shahreza, M. Multilingual CAPTCHA. In: 2007 IEEE Int. Conf. on Computational Cybernetics, pp. 135\u0026ndash;139. (2007).\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eShirali-Shahreza, M. H. \u0026amp; Shirali-Shahreza, M. Advanced Nastaliq CAPTCHA. In: 7th IEEE Int. Conf. on Cybernetic Intelligent Systems, pp. 1\u0026ndash;3. (2008).\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eAlsuhibany, S. A. \u0026amp; Alquraishi, M. Usability and Security of Arabic Text-based CAPTCHA Using Visual Cryptography. \u003cem\u003eComput. Syst. Sci. Eng.\u003c/em\u003e \u003cb\u003e40\u003c/b\u003e (2), 421\u0026ndash;440 (2022).\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eParvez, M. T., Alsuhibani, A. M. \u0026amp; Alamri, A. H. Educational and Cybersecurity Applications of an Arabic CAPTCHA Gamification System. Ing\u0026eacute;nierie des Syst\u0026egrave;mes. \u003cem\u003ed\u0026rsquo;Information\u003c/em\u003e \u003cb\u003e28\u003c/b\u003e (5), 1275 (2023).\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eLajmi, H., Idoudi, F., Njah, H., Kammoun, H. M. \u0026amp; Njah, I. Strengthening Applications' Security with Handwritten Arabic Calligraphy CAPTCHA. In: 8th IEEE Forum on Research and Technologies for Society and Industry Innovation (RTSI), pp. 548\u0026ndash;553. (2024).\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eAldosari, M. H. Innovative multilingual CAPTCHA Based on handwritten characteristics. In: Multilingualism and Bilingualism. IntechOpen. (2017).\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eParvez, M. T. \u0026amp; Alsuhibany, S. A. Segmentation-Validation based Handwritten Arabic CAPTCHA Generation. \u003cem\u003eComput. Secur.\u003c/em\u003e \u003cb\u003e96\u003c/b\u003e, 101829. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.1016/j.cose.2020.101829\u003c/span\u003e\u003cspan address=\"10.1016/j.cose.2020.101829\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e (2020).\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eAldosari, M. H. \u0026amp; Al-Daraiseh, A. A. Strong multilingual CAPTCHA based on handwritten characters. In: 7th Int. Conf. on Information and Communication Systems (ICICS), pp. 239\u0026ndash;245. (2016).\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eFawa\u0026rsquo;reh, M., Qasaimeh, M., AbuArja, I. \u0026amp; Al-Fayoumi, M. Mitigating Deep learning Attacks Against Text Image CAPTCHA Using Arabic Scheme. \u003cem\u003eInt. J. Commun. Antenna Propag.\u003c/em\u003e \u003cb\u003e11\u003c/b\u003e (4), 288\u0026ndash;296 (2021).\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eAlsuhibany, S. A. \u0026amp; Alnoshan, A. A. Interactive Handwritten and Text-Based Handwritten Arabic CAPTCHA Schemes for Mobile Devices: A Comparative Study. \u003cem\u003eIEEE Access.\u003c/em\u003e \u003cb\u003e9\u003c/b\u003e, 140991\u0026ndash;141001 (2021).\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eAbubaker, H., Salah, K., Al-Muhairi, H. \u0026amp; Bentiba, A. Digital Arabic content: challenges and opportunities. In: Int. Conf. on ICT Research (ICTRC), pp. 330\u0026ndash;333. (2015).\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eAbubaker, H., Salah, K., Al-Muhairi, H. \u0026amp; Bentiba, A. Arabic reCAPTCHA Service for Enhancing Digitization of Arabic Manuscripts. \u003cem\u003eArab. J. Sci. Eng.\u003c/em\u003e \u003cb\u003e42\u003c/b\u003e (8), 3391\u0026ndash;3408 (2017).\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eAbubaker, H. B. Arabic reCAPTCHA Service for Web Security and Old Arabic Manuscripts Digitization. Doctoral Dissertation, Khalifa Univ. Sci. Technol. Res. (2016).\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eBirch, J. \u003cem\u003eDiagnosis of Defective Colour Vision\u003c/em\u003e second edn (Oxford University Press, 2012).\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eCole, B. L. Assessment of inherited colour vision defects in clinical practice. \u003cem\u003eClin. Exp. Optom.\u003c/em\u003e \u003cb\u003e87\u003c/b\u003e (4\u0026ndash;5), 276\u0026ndash;293 (2004).\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eHurvich, L. M. \u0026amp; Jameson, D. An opponent-process theory of color vision. \u003cem\u003ePsychol. Rev.\u003c/em\u003e \u003cb\u003e64\u003c/b\u003e (6), 384\u0026ndash;404 (1957).\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eIshihara, S. Tests for Colour Blindness. Kanehara Shuppan (1972).\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eHardy, G. H., Rand, G. \u0026amp; Rittler, M. C. HRR Pseudoisochromatic Plates. Richmond Products. (1954).\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eMollon, J. D. \u0026amp; Reffin, J. P. The Cambridge Colour Test. \u003cem\u003eJ. Opt. Soc. Am. A\u003c/em\u003e. \u003cb\u003e6\u003c/b\u003e (3), 586\u0026ndash;597 (1989).\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eAlsuhibany, S. A., Almohaimeed, F. N. \u0026amp; Alrobah, N. A. Synthetic Arabic handwritten CAPTCHA. \u003cem\u003eInt. J. Inf. Comput. Secur.\u003c/em\u003e \u003cb\u003e16\u003c/b\u003e (3\u0026ndash;4), 385\u0026ndash;398 (2021).\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eBursztein, E. et al. Easy does it: More usable CAPTCHAs. In: Proc. SIGCHI Conf. Hum. Factors Comput. Syst., pp. 2637\u0026ndash;2646. (2014).\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eGoogle Cloud Cloud Vision API. (2022). \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://cloud.google.com/vision\u003c/span\u003e\u003cspan address=\"https://cloud.google.com/vision\" targettype=\"URL\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e (accessed 10 June 2025).\u003c/span\u003e\u003c/li\u003e\u003c/ol\u003e"}],"fulltextSource":"","fullText":"","funders":[],"hasAdminPriorityOnWorkflow":false,"hasManuscriptDocX":true,"hasOptedInToPreprint":true,"hasPassedJournalQc":"","hasAnyPriority":false,"hideJournal":true,"highlight":"","institution":"","isAcceptedByJournal":false,"isAuthorSuppliedPdf":false,"isDeskRejected":"","isHiddenFromSearch":false,"isInQc":false,"isInWorkflow":false,"isPdf":false,"isPdfUpToDate":true,"isWithdrawnOrRetracted":false,"journal":{"display":true,"email":"[email protected]","identity":"researchsquare","isNatureJournal":false,"hasQc":true,"allowDirectSubmit":true,"externalIdentity":"","sideBox":"","snPcode":"","submissionUrl":"/submission","title":"Research Square","twitterHandle":"researchsquare","acdcEnabled":true,"dfaEnabled":false,"editorialSystem":"","reportingPortfolio":"","inReviewEnabled":false,"inReviewRevisionsEnabled":true},"keywords":"Security, Arabic CAPTCHA, Usability, Pseudoisochromatic Plates, Experimental study, AI-Resistant CAPTCHA","lastPublishedDoi":"10.21203/rs.3.rs-7367912/v1","lastPublishedDoiUrl":"https://doi.org/10.21203/rs.3.rs-7367912/v1","license":{"name":"CC BY 4.0","url":"https://creativecommons.org/licenses/by/4.0/"},"manuscriptAbstract":"\u003cp\u003eThis paper presents a novel Arabic CAPTCHA generation framework that leverages Pseudoisochromatic Plates (PIPs)\u0026mdash;a technique traditionally used in color vision testing\u0026mdash;to enhance security and usability in human verification systems. The proposed scheme embeds both printed and handwritten Arabic texts into PIP-style backgrounds using randomized color blending and non-font-based character rendering. This design introduces high visual complexity that resists segmentation and recognition by automated solvers while maintaining legibility for human users. A comprehensive experimental evaluation involving human participants and machine recognition tools demonstrates the robustness of the proposed approach against state-of-the-art attacks, including Google Vision API. Usability assessments further indicate that, while handwritten text embedded in PIP backgrounds achieves superior security, printed text offers better efficiency and effectiveness. The results underscore a critical trade-off between CAPTCHA security and usability and highlight the potential of PIP-based mechanisms to improve Arabic CAPTCHA systems. This work introduces a new direction in CAPTCHA design tailored for linguistically complex scripts and provides valuable insights for developing secure and user-inclusive verification tools in Arabic-speaking digital environments.\u003c/p\u003e","manuscriptTitle":"A Novel Arabic CAPTCHA Generation Framework Based on Pseudoisochromatic Plates for Enhanced Human Verification","msid":"","msnumber":"","nonDraftVersions":[{"code":1,"date":"2025-09-23 18:11:08","doi":"10.21203/rs.3.rs-7367912/v1","editorialEvents":[{"type":"communityComments","content":0}],"status":"published","journal":{"display":true,"email":"[email protected]","identity":"researchsquare","isNatureJournal":false,"hasQc":true,"allowDirectSubmit":true,"externalIdentity":"","sideBox":"","snPcode":"","submissionUrl":"/submission","title":"Research Square","twitterHandle":"researchsquare","acdcEnabled":true,"dfaEnabled":false,"editorialSystem":"","reportingPortfolio":"","inReviewEnabled":false,"inReviewRevisionsEnabled":true}}],"origin":"","ownerIdentity":"d2bf8669-c519-4c0e-b14c-b5b793a187c6","owner":[],"postedDate":"September 23rd, 2025","published":true,"recentEditorialEvents":[],"rejectedJournal":[],"revision":"","amendment":"","status":"posted","subjectAreas":[{"id":55050513,"name":"Physical sciences/Engineering"},{"id":55050514,"name":"Physical sciences/Mathematics and computing"}],"tags":[],"updatedAt":"2025-10-08T05:53:32+00:00","versionOfRecord":[],"versionCreatedAt":"2025-09-23 18:11:08","video":"","vorDoi":"","vorDoiUrl":"","workflowStages":[]},"version":"v1","identity":"rs-7367912","journalConfig":"researchsquare"},"__N_SSP":true},"page":"/article/[identity]/[[...version]]","query":{"redirect":"/article/rs-7367912","identity":"rs-7367912","version":["v1"]},"buildId":"8U1c8b4HqxoKbykW_rLl7","isFallback":false,"isExperimentalCompile":false,"dynamicIds":[84888],"gssp":true,"scriptLoader":[]}