An Artificial Intelligence powered Digital Inline Holographic Microscopy and Characterization Scheme

An Artificial Intelligence powered Digital Inline Holographic Microscopy and Characterization Scheme | 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 An Artificial Intelligence powered Digital Inline Holographic Microscopy and Characterization Scheme Rajkumar Vaghashiya, Varun Chauhan, Kaushal Kapadiya, Smit Sanghavi, and 6 more This is a preprint; it has not been peer reviewed by a journal. https://doi.org/ 10.21203/rs.3.rs-51993/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 Digital Inline Holography (DIH) based microscopy is a well-established technique for the characterization of nano and microparticles, such as biological cells, artificial microparticles, quantum dots, etc. Due to its simplicity and cost-effectiveness, various practical solutions such as auto characterization of complete blood count (CBC), cell viability test, and 3D cell tomography have been developed. In our previous work, we demonstrated the feasibility of this system to perform complete blood count along with the auto characterization of cell-lines as well as shape and size characterization of the microparticles. However, its performance suffered due to the weak signals from some of the cells owing to their poor signatures and the presence of background noise. The auto characterization technique therein was based on the parameters determined from our empirical findings, which limit the system in terms of its cellline recognition power. In this work, we try to address these issues by leveraging an artificial intelligence-powered auto signal enhancing scheme as well as adaptive cell characterization technique. The performance comparison of our proposed method with the existing analytical model shows an increase in accuracy to >98% along with the signal enhancement of >5 dB for most cell types like Red Blood Cell (RBC) and White Blood Cell (WBC), except the cancer cells (HepG2 and MCF-7) for which the accuracy is about 84%. Biomedical Engineering Digital Inline Holography DIH auto characterization technique artificial intelligence Figures Figure 1 Figure 2 Figure 3 Figure 4 Figure 5 Figure 6 Full Text Additional Declarations There is NO Competing Interest. Supplementary Files Supplementary.pdf Supplementary file 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-51993","acceptedTermsAndConditions":true,"allowDirectSubmit":true,"archivedVersions":[],"articleType":"Article","associatedPublications":[],"authors":[{"id":1143396,"identity":"f5cc4047-08db-466f-aa1c-564b37544682","order_by":0,"name":"Rajkumar Vaghashiya","email":"","orcid":"https://orcid.org/0000-0001-6621-4106","institution":"PDPU","correspondingAuthor":false,"submittingAuthor":false,"prefix":"","firstName":"Rajkumar","middleName":"","lastName":"Vaghashiya","suffix":""},{"id":1143397,"identity":"87542f77-8ab6-4b2f-b939-eb4de1cec5aa","order_by":1,"name":"Varun Chauhan","email":"","orcid":"","institution":"PDPU","correspondingAuthor":false,"submittingAuthor":false,"prefix":"","firstName":"Varun","middleName":"","lastName":"Chauhan","suffix":""},{"id":1143398,"identity":"abe299fb-d341-461c-b326-2ccc4f9248de","order_by":2,"name":"Kaushal Kapadiya","email":"","orcid":"","institution":"PDPU","correspondingAuthor":false,"submittingAuthor":false,"prefix":"","firstName":"Kaushal","middleName":"","lastName":"Kapadiya","suffix":""},{"id":1143399,"identity":"4809d121-4c1f-4de7-9890-bfdb09cf3fe3","order_by":3,"name":"Smit Sanghavi","email":"","orcid":"","institution":"PDPU","correspondingAuthor":false,"submittingAuthor":false,"prefix":"","firstName":"Smit","middleName":"","lastName":"Sanghavi","suffix":""},{"id":1143400,"identity":"333ac43a-8e19-43eb-947b-f62e4f88e851","order_by":4,"name":"Ishita Nandwani","email":"","orcid":"","institution":"PDPU","correspondingAuthor":false,"submittingAuthor":false,"prefix":"","firstName":"Ishita","middleName":"","lastName":"Nandwani","suffix":""},{"id":1143401,"identity":"0863eae0-3eaf-4229-8cbe-57b8b8908988","order_by":5,"name":"Jeanie Basumatary","email":"","orcid":"","institution":"PDPU","correspondingAuthor":false,"submittingAuthor":false,"prefix":"","firstName":"Jeanie","middleName":"","lastName":"Basumatary","suffix":""},{"id":1143402,"identity":"cbb1a6df-0734-462e-84cd-ca3f5ab757c6","order_by":6,"name":"Riya Thakore","email":"","orcid":"","institution":"PDPU","correspondingAuthor":false,"submittingAuthor":false,"prefix":"","firstName":"Riya","middleName":"","lastName":"Thakore","suffix":""},{"id":1143403,"identity":"8801c379-09f1-4f09-8334-2f9f10dc06ca","order_by":7,"name":"Dongmin Seo","email":"","orcid":"","institution":"Korea Research Institute of Ships \u0026 Ocean Engineering","correspondingAuthor":false,"submittingAuthor":false,"prefix":"","firstName":"Dongmin","middleName":"","lastName":"Seo","suffix":""},{"id":1143404,"identity":"a2aa133f-4026-45eb-931c-8be9359fbf2c","order_by":8,"name":"Sungkyu Seo","email":"","orcid":"","institution":"Korea University","correspondingAuthor":false,"submittingAuthor":false,"prefix":"","firstName":"Sungkyu","middleName":"","lastName":"Seo","suffix":""},{"id":1143405,"identity":"ed9fce8b-70f1-46ce-979d-4651362f38e5","order_by":9,"name":"Mohendra Roy","email":"data:image/png;base64,iVBORw0KGgoAAAANSUhEUgAAAZAAAAAyAQMAAABI0h/eAAAABlBMVEX///8AAABVwtN+AAAACXBIWXMAAA7EAAAOxAGVKw4bAAAA4klEQVRIiWNgGAWjYFACHiBmg7I/GJCqhXEGyVqYeYjRoNvee/BxQdk9eQbpw8ekbQps5PgZmI99/IJHi9mZc8nGM84VGzbwpaVJ5xikGUs2sCXPlsGn5UaOmTRvWwJjAw+PGVDL4cQNB3iMmSXwazH/DdRiD9ZiYfC/fj8RWsyYgVoSwVoYDA4kGDDwGDN+wOuXM8bSPOcSktt42JItewySDWccZktmxqODwex4j+FnnrIE234e5oM3fvyxk+dvbz7M+AOfHhgARg0LxAfMxEYQSC3cB8TZMgpGwSgYBSMFAAAO4kDxTNSU+gAAAABJRU5ErkJggg==","orcid":"https://orcid.org/0000-0001-5815-3294","institution":"Pandit Deendayal Petroleum University","correspondingAuthor":true,"submittingAuthor":false,"prefix":"","firstName":"Mohendra","middleName":"","lastName":"Roy","suffix":""}],"badges":[],"createdAt":"2020-07-31 19:31:35","currentVersionCode":1,"declarations":"","doi":"10.21203/rs.3.rs-51993/v1","doiUrl":"https://doi.org/10.21203/rs.3.rs-51993/v1","draftVersion":[],"editorialEvents":[],"editorialNote":"","failedWorkflow":false,"files":[{"id":1784968,"identity":"39b499f7-ce5a-4737-be5b-212f9100b5ef","added_by":"auto","created_at":"2020-08-04 21:51:59","extension":"png","order_by":1,"title":"Figure 1","display":"","copyAsset":false,"role":"figure","size":884917,"visible":true,"origin":"","legend":"Schematic of the DIH microscope setup and the proposed neural network architecture for the auto characterization of DIH micrographs. (a) schematic of the principle of diffraction signature generation of a micro object, (b) schematic of the DIH imaging setup which shows the simplicity of the setup, (c) schematic of the dataset creation process by cropping individual cell diffraction pattern from the whole frame DIH micrograph, (d) the schematic of the proposed denoising and classification architecture. Here, the denoising autoencoder enhances the signal of the individual cells which is then fed to the CNN module for classification.","description":"","filename":"F1.png","url":"https://assets-eu.researchsquare.com/files/rs-51993/v1/F1.png"},{"id":1784969,"identity":"b3daa0d3-cd1a-43ec-8e2c-e03903ece7cf","added_by":"auto","created_at":"2020-08-04 21:51:59","extension":"png","order_by":2,"title":"Figure 2","display":"","copyAsset":false,"role":"figure","size":470236,"visible":true,"origin":"","legend":"Reconstructed results from the optimized CNN. The top row is the original DIH image of a single RBC, WBC, MCF7, HepG2, 10µm bead, and 20µm bead. The middle row is the noisy version (with variance 100) of the corresponding original images. The third row is the reconstructed version of the corresponding original images from the noisy image. The fourth row is the 2D intensity contour plot of the original image to show the unique signature of each of these cell lines.","description":"","filename":"F2.png","url":"https://assets-eu.researchsquare.com/files/rs-51993/v1/F2.png"},{"id":1784970,"identity":"1a49b82c-b6db-4c91-ad5c-c8fe62f4938c","added_by":"auto","created_at":"2020-08-04 21:51:59","extension":"png","order_by":3,"title":"Figure 3","display":"","copyAsset":false,"role":"figure","size":877248,"visible":true,"origin":"","legend":"Results from the CNN and ELM autoencoder. (a) The performance of the CNN autoencoder for improved SNR (average of all the classes) across various layers and kernel sizes. (b) The performance of the CNN autoencoder across varying noise levels. Here, the variance of the Gaussian noise ranges from 100 to 600, and evaluated on the optimal network architecture i.e. 3,3,5,5,7,7,1. (c) The performance of the CNN autoencoder across various input sizes (cropping size). Here, the sizes vary from 66 × 66 to 36 × 36. (d) The convergence of the optimal CNN network with the number of samples for the first epoch. (e) The number of hidden layer neurons in ELM autoencoder vs improvement in SNR, (f) Variance vs improved SNR for ELM autoencoder, (g) Input size vs improved SNR for ELM autoencoder, and (h) The convergence of ELM autoencoder within the 1st epoch.","description":"","filename":"F3.png","url":"https://assets-eu.researchsquare.com/files/rs-51993/v1/F3.png"},{"id":1784971,"identity":"07651eab-08e3-4a0e-9f45-7e7ae33dac18","added_by":"auto","created_at":"2020-08-04 21:51:59","extension":"png","order_by":4,"title":"Figure 4","display":"","copyAsset":false,"role":"figure","size":234695,"visible":true,"origin":"","legend":"The CNN Architectures with varying depth. The models in the solid box are for the optimization of the model with varying depth. The models inside the dotted box are for the optimization of the parameters. Here the orange line is the input layer and the rectangle represent the 2D convolution layer. The kernel size and number of kernels are indicated inside the round brackets, the dropout rate in square brackets, and the number of neurons in fully connected layers are placed directly within the rectangle. The aqua blue, green and purple lines are the max-pool layers. The red line is the output layer and uses a SoftMax activation function. The final optimized architecture is as represented by the last 3D figure.","description":"","filename":"F4.png","url":"https://assets-eu.researchsquare.com/files/rs-51993/v1/F4.png"},{"id":1784972,"identity":"47fd3c92-5a66-44c8-bfb2-1d956cdcde5d","added_by":"auto","created_at":"2020-08-04 21:51:59","extension":"png","order_by":5,"title":"Figure 5","display":"","copyAsset":false,"role":"figure","size":686003,"visible":true,"origin":"","legend":"Results for the optimization of the CNN model. (a-c) shows the performance of shallow, intermediate, and deep networks respectively. (d-g) Performance of the fine-tuned intermediate architecture. (h) The confusion matrix, showing classification accuracy on the test dataset across all cell lines using the fine-tuned and optimized model. (i) The receiver operating characteristic (ROC) curve for each of the cell lines for the optimized model.","description":"","filename":"F5.png","url":"https://assets-eu.researchsquare.com/files/rs-51993/v1/F5.png"},{"id":1784973,"identity":"d7e30ed2-f041-43a0-b5e4-d13bb5477808","added_by":"auto","created_at":"2020-08-04 21:52:00","extension":"png","order_by":6,"title":"Figure 6","display":"","copyAsset":false,"role":"figure","size":258258,"visible":true,"origin":"","legend":"The results from the transfer of learning study. (a) epoch vs. accuracy graph for the initial epochs in the initial training phase without RBC and then the transfer of training with RBC. The blue and saffron colour represent the initial training accuracy and validation curves. The green and red lines represent the transfer learning accuracy and validation curves, (b) epoch vs. loss, (c) the confusion matrix from the pre-trained model,\n(d) the confusion matrix with the re-trained model.\n","description":"","filename":"F6.png","url":"https://assets-eu.researchsquare.com/files/rs-51993/v1/F6.png"},{"id":13520195,"identity":"82181092-77d3-425b-8d09-55b8fdf28023","added_by":"auto","created_at":"2021-09-17 00:26:32","extension":"pdf","order_by":1,"title":"","display":"","copyAsset":false,"role":"manuscript-pdf","size":1656949,"visible":true,"origin":"","legend":"","description":"","filename":"ANNDIHMohendra.pdf","url":"https://assets-eu.researchsquare.com/files/rs-51993/v1_covered.pdf"},{"id":1784978,"identity":"a0980d1d-e5a4-42ff-8fea-0c85d89cf66a","added_by":"auto","created_at":"2020-08-04 21:52:15","extension":"pdf","order_by":1,"title":"","display":"","copyAsset":false,"role":"manuscript-pdf","size":1564049,"visible":true,"origin":"","legend":"","description":"","filename":"ANNDIHMohendra.pdf","url":"https://assets-eu.researchsquare.com/files/rs-51993/v1_stamped.pdf"},{"id":1784974,"identity":"90874366-e00a-4514-b6a1-47775d777760","added_by":"auto","created_at":"2020-08-04 21:52:00","extension":"pdf","order_by":1,"title":"","display":"","copyAsset":false,"role":"manuscript-pdf","size":1511718,"visible":true,"origin":"","legend":"","description":"","filename":"ANNDIHMohendra.pdf","url":"https://assets-eu.researchsquare.com/files/rs-51993/v1/ANNDIHMohendra.pdf"},{"id":1784975,"identity":"554adab1-a2ce-437a-9f08-e3ee7218e09c","added_by":"auto","created_at":"2020-08-04 21:52:00","extension":"pdf","order_by":1,"title":"","display":"","copyAsset":false,"role":"supplement","size":218824,"visible":true,"origin":"","legend":"Supplementary file","description":"","filename":"Supplementary.pdf","url":"https://assets-eu.researchsquare.com/files/rs-51993/v1/Supplementary.pdf"}],"financialInterests":"There is \u003cb\u003eNO\u003c/b\u003e Competing Interest.","formattedTitle":"\u003cp\u003eAn Artificial Intelligence powered Digital Inline Holographic Microscopy and Characterization Scheme\u003c/p\u003e","fulltext":[{"header":"Full Text","content":"\u003cp\u003eThis preprint is available for \u003ca href='/article/rs-51993/latest.pdf' target='_blank'\u003edownload as a PDF\u003c/a\u003e.\u003c/p\u003e"}],"fulltextSource":"","fullText":"","funders":[],"hasAdminPriorityOnWorkflow":false,"hasManuscriptDocX":false,"hasOptedInToPreprint":true,"hasPassedJournalQc":"","hasAnyPriority":true,"hideJournal":true,"highlight":"","institution":"","isAcceptedByJournal":false,"isAuthorSuppliedPdf":true,"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":"Digital Inline Holography, DIH, auto characterization technique, artificial intelligence","lastPublishedDoi":"10.21203/rs.3.rs-51993/v1","lastPublishedDoiUrl":"https://doi.org/10.21203/rs.3.rs-51993/v1","license":{"name":"CC BY 4.0","url":"https://creativecommons.org/licenses/by/4.0/"},"manuscriptAbstract":"Digital Inline Holography (DIH) based microscopy is a well-established technique for the characterization of nano and microparticles, such as biological cells, artificial microparticles, quantum dots, etc. Due to its simplicity and cost-effectiveness, various practical solutions such as auto characterization of complete blood count (CBC), cell viability test, and 3D cell tomography have been developed. In our previous work, we demonstrated the feasibility of this system to perform complete blood count along with the auto characterization of cell-lines as well as shape and size characterization of the microparticles. However, its performance suffered due to the weak signals from some of the cells owing to their poor signatures and the presence of background noise. The auto characterization technique therein was based on the parameters determined from our empirical findings, which limit the system in terms of its cellline recognition power. In this work, we try to address these issues by leveraging an artificial intelligence-powered auto signal enhancing scheme as well as adaptive cell characterization technique. The performance comparison of our proposed method with the existing analytical model shows an increase in accuracy to \u003e98% along with the signal enhancement of \u003e5 dB for most cell types like Red Blood Cell (RBC) and White Blood Cell (WBC), except the cancer cells (HepG2 and MCF-7) for which the accuracy is about 84%.","manuscriptTitle":"An Artificial Intelligence powered Digital Inline Holographic Microscopy and Characterization Scheme","msid":"","msnumber":"","nonDraftVersions":[{"code":1,"date":"2020-08-04 21:48:08","doi":"10.21203/rs.3.rs-51993/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":"bd5fd457-9331-48b1-ab58-aca445a77896","owner":[],"postedDate":"August 4th, 2020","published":true,"recentEditorialEvents":[],"rejectedJournal":[],"revision":"","amendment":"","status":"posted","subjectAreas":[{"id":242137,"name":"Biomedical Engineering"}],"tags":[],"updatedAt":"2020-08-04T21:48:08+00:00","versionOfRecord":[],"versionCreatedAt":"2020-08-04 21:48:08","video":"","vorDoi":"","vorDoiUrl":"","workflowStages":[]},"version":"v1","identity":"rs-51993","journalConfig":"researchsquare"},"__N_SSP":true},"page":"/article/[identity]/[[...version]]","query":{"redirect":"/article/rs-51993","identity":"rs-51993","version":["v1"]},"buildId":"cBFmMYwuxLRRLfASyISRj","isFallback":false,"isExperimentalCompile":false,"dynamicIds":[84888],"gssp":true,"scriptLoader":[]}