A Spectrogram and Local Feature-Assisted Convolutional Neural Network for Amharic Speech Emotion Identification

A Spectrogram and Local Feature-Assisted Convolutional Neural Network for Amharic Speech Emotion Identification | 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 Spectrogram and Local Feature-Assisted Convolutional Neural Network for Amharic Speech Emotion Identification Yeshambel Asmare Mengist, Abrham Debasu Mengistu, Mulatu Yirga Beyene, and 2 more This is a preprint; it has not been peer reviewed by a journal. https://doi.org/ 10.21203/rs.3.rs-8961140/v1 This work is licensed under a CC BY 4.0 License Status: Under Review Version 1 posted 10 You are reading this latest preprint version Abstract Speech Emotion Recognition (SER) plays a significant role in improving human-computer interaction and human-human communication. Nevertheless, speech emotion recognition in low-resource languages like Amharic is still a difficult task because of the lack of datasets and language diversity. In this paper, a Convolutional Neural Network (CNN)-based approach, which combines spectrogram features and local acoustic features such as Mel-Frequency Cepstral Coefficients (MFCCs), chroma, zero-crossing rate (ZCR), energy, and pitch, is proposed for efficient Amharic speech emotion recognition. A dataset of 1,650 three-second Amharic speech samples was created, and the samples were labeled with five emotional classes: anger, fear, happy, neutral, and sad. Advanced preprocessing methods such as spectral subtraction and wavelet denoising were used to improve the quality of the signals and speed up the training process. The experimental results show that the proposed CNN-based approach has a classification accuracy of 90 percent, which is better than the recurrent neural network-based approaches: Long Short-Term Memory (LSTM) with 58.48 percent, Bidirectional Long Short-Term Memory (BiLSTM) with 63.33 percent, and Gated Recurrent Unit (GRU) with 40 percent, as well as the single-feature models: local acoustic features with 73 percent and spectrogram features with 79 percent. These results confirm that the integration of spectrogram and local acoustic features within a CNN architecture improves accuracy and efficiency in speech emotion recognition in low-resource languages, setting a standard for future Amharic SER research. Biological sciences/Computational biology and bioinformatics Physical sciences/Engineering Physical sciences/Mathematics and computing Biological sciences/Neuroscience Speech Emotion Identification Convolutional Neural Network Spectrogram Local Features Amharic Speech Deep Learning Full Text Additional Declarations No competing interests reported. Cite Share Download PDF Status: Under Review Version 1 posted Editorial decision: Revision requested 30 Mar, 2026 Reviews received at journal 30 Mar, 2026 Reviews received at journal 24 Mar, 2026 Reviewers agreed at journal 24 Mar, 2026 Reviewers agreed at journal 20 Mar, 2026 Reviewers invited by journal 20 Mar, 2026 Editor assigned by journal 20 Mar, 2026 Editor invited by journal 17 Mar, 2026 Submission checks completed at journal 14 Mar, 2026 First submitted to journal 13 Mar, 2026 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-8961140","acceptedTermsAndConditions":true,"allowDirectSubmit":false,"archivedVersions":[],"articleType":"Article","associatedPublications":[],"authors":[{"id":611431512,"identity":"99212613-d5f3-4df3-bd62-bcecd57a694a","order_by":0,"name":"Yeshambel Asmare Mengist","email":"data:image/png;base64,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","orcid":"","institution":"Debark University","correspondingAuthor":true,"prefix":"","firstName":"Yeshambel","middleName":"Asmare","lastName":"Mengist","suffix":""},{"id":611431513,"identity":"03df0ce1-6025-4236-83b6-74eac7db7ea3","order_by":1,"name":"Abrham Debasu Mengistu","email":"","orcid":"","institution":"Bahir Dar University","correspondingAuthor":false,"prefix":"","firstName":"Abrham","middleName":"Debasu","lastName":"Mengistu","suffix":""},{"id":611431514,"identity":"d37f7547-b132-4981-869c-9ed0290eabf4","order_by":2,"name":"Mulatu Yirga Beyene","email":"","orcid":"","institution":"Wollo University","correspondingAuthor":false,"prefix":"","firstName":"Mulatu","middleName":"Yirga","lastName":"Beyene","suffix":""},{"id":611431515,"identity":"e494f88c-abce-4ec2-94f9-457b1b3214d5","order_by":3,"name":"Mikiyas Assefa Kassa","email":"","orcid":"","institution":"Debark University","correspondingAuthor":false,"prefix":"","firstName":"Mikiyas","middleName":"Assefa","lastName":"Kassa","suffix":""},{"id":611431516,"identity":"1a1dd021-f273-4a55-9d3f-7670eaae506f","order_by":4,"name":"Getasew Asmare Mengist","email":"","orcid":"","institution":"Bahir Dar University","correspondingAuthor":false,"prefix":"","firstName":"Getasew","middleName":"Asmare","lastName":"Mengist","suffix":""}],"badges":[],"createdAt":"2026-02-24 21:23:43","currentVersionCode":1,"declarations":"","doi":"10.21203/rs.3.rs-8961140/v1","doiUrl":"https://doi.org/10.21203/rs.3.rs-8961140/v1","draftVersion":[],"editorialEvents":[],"editorialNote":"","failedWorkflow":false,"files":[{"id":105565764,"identity":"0a0d98a2-8a70-40e9-b4b0-bfc314e4dd6d","added_by":"auto","created_at":"2026-03-27 12:54:20","extension":"pdf","order_by":1,"title":"","display":"","copyAsset":false,"role":"manuscript-pdf","size":1069201,"visible":true,"origin":"","legend":"","description":"","filename":"ASEIUpdated.pdf","url":"https://assets-eu.researchsquare.com/files/rs-8961140/v1_covered_5ca12540-8587-4791-88bd-5955fe8d3fea.pdf"}],"financialInterests":"No competing interests reported.","formattedTitle":"A Spectrogram and Local Feature-Assisted Convolutional Neural Network for Amharic Speech Emotion Identification","fulltext":[],"fulltextSource":"","fullText":"","funders":[],"hasAdminPriorityOnWorkflow":false,"hasManuscriptDocX":false,"hasOptedInToPreprint":true,"hasPassedJournalQc":"","hasAnyPriority":false,"hideJournal":false,"highlight":"","institution":"","isAcceptedByJournal":true,"isAuthorSuppliedPdf":true,"isDeskRejected":"","isHiddenFromSearch":false,"isInQc":false,"isInWorkflow":false,"isPdf":true,"isPdfUpToDate":true,"isWithdrawnOrRetracted":false,"journal":{"display":true,"email":"[email protected]","identity":"scientific-reports","isNatureJournal":false,"hasQc":true,"allowDirectSubmit":false,"externalIdentity":"scirep","sideBox":"Learn more about [Scientific Reports](http://www.nature.com/srep/)","snPcode":"","submissionUrl":"","title":"Scientific Reports","twitterHandle":"","acdcEnabled":true,"dfaEnabled":true,"editorialSystem":"stoa","reportingPortfolio":"Scientific Reports","inReviewEnabled":true,"inReviewRevisionsEnabled":true},"keywords":"Speech Emotion Identification, Convolutional Neural Network, Spectrogram, Local Features, Amharic Speech, Deep Learning","lastPublishedDoi":"10.21203/rs.3.rs-8961140/v1","lastPublishedDoiUrl":"https://doi.org/10.21203/rs.3.rs-8961140/v1","license":{"name":"CC BY 4.0","url":"https://creativecommons.org/licenses/by/4.0/"},"manuscriptAbstract":"\u003cp\u003eSpeech Emotion Recognition (SER) plays a significant role in improving human-computer interaction and human-human communication. Nevertheless, speech emotion recognition in low-resource languages like Amharic is still a difficult task because of the lack of datasets and language diversity. In this paper, a Convolutional Neural Network (CNN)-based approach, which combines spectrogram features and local acoustic features such as Mel-Frequency Cepstral Coefficients (MFCCs), chroma, zero-crossing rate (ZCR), energy, and pitch, is proposed for efficient Amharic speech emotion recognition. A dataset of 1,650 three-second Amharic speech samples was created, and the samples were labeled with five emotional classes: anger, fear, happy, neutral, and sad. Advanced preprocessing methods such as spectral subtraction and wavelet denoising were used to improve the quality of the signals and speed up the training process. The experimental results show that the proposed CNN-based approach has a classification accuracy of 90 percent, which is better than the recurrent neural network-based approaches: Long Short-Term Memory (LSTM) with 58.48 percent, Bidirectional Long Short-Term Memory (BiLSTM) with 63.33 percent, and Gated Recurrent Unit (GRU) with 40 percent, as well as the single-feature models: local acoustic features with 73 percent and spectrogram features with 79 percent. These results confirm that the integration of spectrogram and local acoustic features within a CNN architecture improves accuracy and efficiency in speech emotion recognition in low-resource languages, setting a standard for future Amharic SER research.\u003c/p\u003e","manuscriptTitle":"A Spectrogram and Local Feature-Assisted Convolutional Neural Network for Amharic Speech Emotion Identification","msid":"","msnumber":"","nonDraftVersions":[{"code":1,"date":"2026-03-25 12:48:52","doi":"10.21203/rs.3.rs-8961140/v1","editorialEvents":[{"type":"communityComments","content":0},{"type":"decision","content":"Revision requested","date":"2026-03-30T09:08:45+00:00","index":"","fulltext":""},{"type":"editorInvitedReview","content":"","date":"2026-03-30T05:03:58+00:00","index":"hide","fulltext":""},{"type":"editorInvitedReview","content":"","date":"2026-03-24T13:58:08+00:00","index":"hide","fulltext":""},{"type":"reviewerAgreed","content":"270733467023600796378384460189772215166","date":"2026-03-24T13:21:38+00:00","index":"hide","fulltext":""},{"type":"reviewerAgreed","content":"282031214041225536082008230188419256814","date":"2026-03-20T13:50:57+00:00","index":"hide","fulltext":""},{"type":"reviewersInvited","content":"","date":"2026-03-20T13:21:39+00:00","index":"","fulltext":""},{"type":"editorAssigned","content":"","date":"2026-03-20T13:19:59+00:00","index":"","fulltext":""},{"type":"editorInvited","content":"","date":"2026-03-17T14:07:17+00:00","index":"","fulltext":""},{"type":"checksComplete","content":"","date":"2026-03-14T06:11:37+00:00","index":"","fulltext":""},{"type":"submitted","content":"Scientific Reports","date":"2026-03-13T17:37:31+00:00","index":"","fulltext":""}],"status":"published","journal":{"display":true,"email":"[email protected]","identity":"scientific-reports","isNatureJournal":false,"hasQc":true,"allowDirectSubmit":false,"externalIdentity":"scirep","sideBox":"Learn more about [Scientific Reports](http://www.nature.com/srep/)","snPcode":"","submissionUrl":"","title":"Scientific Reports","twitterHandle":"","acdcEnabled":true,"dfaEnabled":true,"editorialSystem":"stoa","reportingPortfolio":"Scientific Reports","inReviewEnabled":true,"inReviewRevisionsEnabled":true}}],"origin":"","ownerIdentity":"5d622136-cfdd-4065-996a-862f99ad7b01","owner":[],"postedDate":"March 25th, 2026","published":true,"recentEditorialEvents":[],"rejectedJournal":[],"revision":"","amendment":"","status":"under-review","subjectAreas":[{"id":65057067,"name":"Biological sciences/Computational biology and bioinformatics"},{"id":65057068,"name":"Physical sciences/Engineering"},{"id":65057069,"name":"Physical sciences/Mathematics and computing"},{"id":65057070,"name":"Biological sciences/Neuroscience"}],"tags":[],"updatedAt":"2026-05-15T06:54:43+00:00","versionOfRecord":[],"versionCreatedAt":"2026-03-25 12:48:52","video":"","vorDoi":"","vorDoiUrl":"","workflowStages":[]},"version":"v1","identity":"rs-8961140","journalConfig":"researchsquare"},"__N_SSP":true},"page":"/article/[identity]/[[...version]]","query":{"redirect":"/article/rs-8961140","identity":"rs-8961140","version":["v1"]},"buildId":"XKTyCvWXoU3ODBz1xrDgd","isFallback":false,"isExperimentalCompile":false,"dynamicIds":[84888],"gssp":true,"scriptLoader":[]}