{"status":"ok","message-type":"work","message-version":"1.0.0","message":{"indexed":{"date-parts":[[2026,8,24]],"date-time":"2026-08-24T18:24:12Z","timestamp":1787595852647,"version":"build-2736575974"},"reference-count":54,"publisher":"Public Library of Science (PLoS)","issue":"8","license":[{"start":{"date-parts":[[2026,8,24]],"date-time":"2026-08-24T00:00:00Z","timestamp":1787529600000},"content-version":"vor","delay-in-days":0,"URL":"http:\/\/creativecommons.org\/licenses\/by\/4.0\/"}],"content-domain":{"domain":["www.plosone.org"],"crossmark-restriction":false},"short-container-title":["PLoS One"],"abstract":"<jats:p>Speech Emotion Recognition (SER) is an important component of human-centered intelligent systems, yet robust performance remains challenging when speaker identities differ between training and testing. This study presents a protocol-aware and reproducible comparison on the CREMA-D corpus using three pipelines: (i) a classical MFCC-based Support Vector Machine (SVM), (ii) a log-mel Convolutional Neural Network (CNN), and (iii) a lightweight hybrid model that concatenates handcrafted acoustic descriptors with CNN-derived embeddings and uses an SVM classifier. The methodological contribution is not a new standalone classifier; it is the controlled integration of identical preprocessing, random and actor-wise evaluation, five-seed robustness reporting, class-wise error analysis, and CPU-oriented deployment within one experimental framework. The Hybrid approach achieves the best overall performance, obtaining 62.03%\u2009\u00b1\u20090.94% Macro-F1 on the random split and 58.09%\u2009\u00b1\u20091.36% on the actor-wise split, outperforming MFCC+SVM (55.72%\u2009\u00b1\u20090.95% and 52.05%\u2009\u00b1\u20091.95%) and the log-mel CNN (50.20%\u2009\u00b1\u20091.51% and 42.68%\u2009\u00b1\u20091.84%). A Streamlit interface supports WAV upload, live prediction, and export of per-seed confusion matrices and summary figures. The results show that a controlled lightweight fusion framework can improve robustness while making the performance gap between speaker-overlapping and speaker-independent evaluation explicit.<\/jats:p>","DOI":"10.1371\/journal.pone.0355238","type":"journal-article","created":{"date-parts":[[2026,8,24]],"date-time":"2026-08-24T17:31:27Z","timestamp":1787592687000},"page":"e0355238","update-policy":"https:\/\/doi.org\/10.1371\/journal.pone.corrections_policy","source":"Crossref","is-referenced-by-count":0,"title":["Hybrid CNN-embedding fusion with MFCC-SVM for speech emotion recognition: Random vs actor-wise evaluation on CREMA-D"],"prefix":"10.1371","volume":"21","author":[{"given":"Parveen","family":"Kumari","sequence":"first","affiliation":[],"role":[{"vocabulary":"crossref","role":"author"}]},{"given":"Yogita Yashveer","family":"Raghav","sequence":"additional","affiliation":[],"role":[{"vocabulary":"crossref","role":"author"}]},{"given":"Vimmi","family":"Kochher","sequence":"additional","affiliation":[],"role":[{"vocabulary":"crossref","role":"author"}]},{"given":"Prashanth kumar","family":"katta","sequence":"additional","affiliation":[],"role":[{"vocabulary":"crossref","role":"author"}]},{"given":"Abhilasha","family":"A","sequence":"additional","affiliation":[],"role":[{"vocabulary":"crossref","role":"author"}]},{"given":"Prabhakar","family":"M","sequence":"additional","affiliation":[],"role":[{"vocabulary":"crossref","role":"author"}]},{"given":"Jayanthiladevi","family":"A","sequence":"additional","affiliation":[],"role":[{"vocabulary":"crossref","role":"author"}]},{"ORCID":"https:\/\/orcid.org\/0009-0004-3400-1324","authenticated-orcid":true,"given":"Fikir Gizachew","family":"Belete","sequence":"additional","affiliation":[],"role":[{"vocabulary":"crossref","role":"author"}]}],"member":"340","published-online":{"date-parts":[[2026,8,24]]},"reference":[{"issue":"5","key":"pone.0355238.ref001","doi-asserted-by":"crossref","first-page":"90","DOI":"10.1145\/3129340","article-title":"Speech emotion recognition: two decades in a nutshell, benchmarks, and ongoing trends","volume":"61","author":"BW Schuller","year":"2018","journal-title":"Commun ACM"},{"issue":"3","key":"pone.0355238.ref002","doi-asserted-by":"crossref","first-page":"572","DOI":"10.1016\/j.patcog.2010.09.020","article-title":"Survey on speech emotion recognition: features, classification schemes, and databases","volume":"44","author":"M El Ayadi 2","year":"2011","journal-title":"Pattern Recognit"},{"issue":"4","key":"pone.0355238.ref003","doi-asserted-by":"crossref","first-page":"377","DOI":"10.1109\/TAFFC.2014.2336244","article-title":"CREMA-D: Crowd-sourced Emotional Multimodal Actors Dataset","volume":"5","author":"H Cao","year":"2014","journal-title":"IEEE Trans Affect Comput"},{"issue":"4","key":"pone.0355238.ref004","doi-asserted-by":"crossref","first-page":"357","DOI":"10.1109\/TASSP.1980.1163420","article-title":"Comparison of parametric representations for monosyllabic word recognition in continuously spoken sentences","volume":"28","author":"S Davis","year":"1980","journal-title":"IEEE Trans Acoust Speech Signal Process"},{"issue":"3","key":"pone.0355238.ref005","doi-asserted-by":"crossref","first-page":"273","DOI":"10.1023\/A:1022627411411","article-title":"Support-vector networks","volume":"20","author":"C Cortes","year":"1995","journal-title":"Mach Learn"},{"key":"pone.0355238.ref006","doi-asserted-by":"crossref","unstructured":"Trigeorgis G, Ringeval F, Brueckner R, Marchi E, Nicolaou MA, Schuller B, et al. Adieu features? End-to-end speech emotion recognition using a deep convolutional recurrent network. In: 2016 IEEE International Conference on Acoustics, Speech and Signal Processing (ICASSP). 2016. p. 5200\u20134.","DOI":"10.1109\/ICASSP.2016.7472669"},{"key":"pone.0355238.ref007","doi-asserted-by":"crossref","unstructured":"Satt A, Rozenberg S, Hoory R. Efficient Emotion Recognition from Speech Using Deep Learning on Spectrograms. In: Proc Interspeech. 2017. p. 1089\u201393.","DOI":"10.21437\/Interspeech.2017-200"},{"key":"pone.0355238.ref008","doi-asserted-by":"crossref","unstructured":"Park DS, Chan W, Zhang Y, Chiu C-C, Zoph B, Cubuk ED, et al. SpecAugment: A Simple Data Augmentation Method for Automatic Speech Recognition. In: Proc Interspeech. 2019. p. 2613\u20137.","DOI":"10.21437\/Interspeech.2019-2680"},{"key":"pone.0355238.ref009","doi-asserted-by":"crossref","unstructured":"Lin T-Y, Goyal P, Girshick R, He K, Dollar P. \u201cFocal loss for dense object detection\u201d. In: Proc. IEEE International Conference on Computer Vision (ICCV). 2017. p. 2980\u20138.","DOI":"10.1109\/ICCV.2017.324"},{"key":"pone.0355238.ref010","unstructured":"Zhang H, Cisse M, Dauphin YN, et al. Mixup: Beyond Empirical Risk Minimization. In: Proceedings of the International Conference on Learning Representations (ICLR). 2018. https:\/\/doi.org\/10.48550\/arXiv.1710.09412"},{"key":"pone.0355238.ref011","doi-asserted-by":"crossref","unstructured":"McFee B, Raffel C, Liang D, Ellis D, McVicar M, Battenberg E, et al. \u201clibrosa: Audio and music signal analysis in Python\u201d. In: Proceedings of the 14th Python in Science Conference (SciPy 2015). 2015. p. 18\u201324.","DOI":"10.25080\/Majora-7b98e3ed-003"},{"key":"pone.0355238.ref012","doi-asserted-by":"crossref","unstructured":"Ibrahim A, Shehata S, Kulkarni A, Mohamed M, Abdul-Mageed M. What Does it Take to Generalize SER Model Across Datasets? A Comprehensive Benchmark. In: Proc Interspeech. 2024. p. 1590\u20134.","DOI":"10.21437\/Interspeech.2024-1983"},{"key":"pone.0355238.ref013","doi-asserted-by":"crossref","first-page":"127015","DOI":"10.1016\/j.neucom.2023.127015","article-title":"A review on speech emotion recognition: A survey, recent advances, challenges, and the influence of noise","volume":"568","author":"SM George","year":"2024","journal-title":"Neurocomputing"},{"key":"pone.0355238.ref014","first-page":"101","article-title":"In defense of one-vs-all classification","volume":"5","author":"R Rifkin","year":"2004","journal-title":"J Mach Learn Res"},{"key":"pone.0355238.ref015","unstructured":"Kingma DP, Ba J. Adam: A method for stochastic optimization. In: International Conference on Learning Representations (ICLR). 2015. https:\/\/doi.org\/10.48550\/arXiv.1412.6980"},{"key":"pone.0355238.ref016","volume-title":"Discrete-Time Signal Processing","author":"AV Oppenheim","year":"1999","edition":"2"},{"key":"pone.0355238.ref017","doi-asserted-by":"crossref","first-page":"13126","DOI":"10.1038\/s41598-024-63776-4","article-title":"An enhanced speech emotion recognition system using vision transformer","volume":"14","author":"O Akinpelu","year":"2024","journal-title":"Sci Rep"},{"key":"pone.0355238.ref018","article-title":"An enhanced speech emotion recognition using vision transformer- based modeling on spectrogram representations","author":"S Akinpelu","year":"2024","journal-title":"Sci Rep"},{"key":"pone.0355238.ref019","article-title":"Enhancing speech emotion recognition through parallel cnns with transformer encoder and co-attention","author":"A Hashem","year":"2025","journal-title":"PeerJ Comput Sci"},{"key":"pone.0355238.ref020","doi-asserted-by":"crossref","unstructured":"Chowdhury JH, et al. \u201cSpeech emotion recognition with light weight deep neural en- semble model using hand crafted features,\u201d Journal\/ venue as listed in PubMed Central, reports very high CREMA-D accuracy under its reported setup; see PMC full text. 2025.","DOI":"10.1038\/s41598-025-95734-z"},{"key":"pone.0355238.ref021","doi-asserted-by":"crossref","unstructured":"Keesing A, Koh YS, Yogarajan V, Witbrock M. \u201cEmotion recognition toolkit (ertk): Standardising tools for emotion recognition research\u201d. In: Proceedings of the 31st ACM International Conference on Multimedia (ACM MM), reports CREMA-D baseline results (e.g., eGeMAPS + RBF-SVM UAR\/Macro-F1). 2023.","DOI":"10.1145\/3581783.3613459"},{"issue":"4","key":"pone.0355238.ref022","doi-asserted-by":"crossref","first-page":"427","DOI":"10.1016\/j.ipm.2009.03.002","article-title":"A systematic analysis of performance measures for classification tasks","volume":"45","author":"M Sokolova","year":"2009","journal-title":"Inf Process Manag"},{"key":"pone.0355238.ref023","unstructured":"Powers DMW. Evaluation: From precision, recall and f-measure to ROC, informed- ness, markedness and correlation. In: Proceedings of the 2011 International Conference on Machine Learning Technologies (MLT 2011). 2011."},{"issue":"2","key":"pone.0355238.ref024","doi-asserted-by":"crossref","first-page":"190","DOI":"10.1109\/TAFFC.2015.2457417","article-title":"The Geneva minimalistic acoustic parameter set (GeMAPS) for voice research and affective computing","volume":"7","author":"F Eyben","year":"2016","journal-title":"IEEE Trans Affect Comput"},{"key":"pone.0355238.ref025","volume-title":"Theory and Applications of Digital Speech Processing","author":"LR Rabiner","year":"2010"},{"key":"pone.0355238.ref026","doi-asserted-by":"crossref","unstructured":"Eyben F, W\u00f6llmer M, Schuller B. Opensmile \u2013 the Munich Versatile and Fast Open-Source Audio Feature Extractor. In: Proceedings of the 18th ACM International Conference on Multimedia. 2010. p. 1459\u201362.","DOI":"10.1145\/1873951.1874246"},{"key":"pone.0355238.ref027","doi-asserted-by":"crossref","unstructured":"Piczak KJ. Esc: dataset for environmental sound classification. In: Proceedings of the 23rd ACM International Conference on Multimedia. 2015. p. 1015\u20138.","DOI":"10.1145\/2733373.2806390"},{"issue":"4","key":"pone.0355238.ref028","doi-asserted-by":"crossref","first-page":"335","DOI":"10.1007\/s10579-008-9076-6","article-title":"IEMOCAP: Interactive emotional dyadic motion capture database","volume":"42","author":"C Busso","year":"2008","journal-title":"Lang Resour Eval"},{"key":"pone.0355238.ref029","article-title":"Speech emotion recognition: Two decades in a nutshell, benchmarks, and ongoing trends","author":"BW Schuller","year":"2018","journal-title":"Commun ACM"},{"issue":"3","key":"pone.0355238.ref030","doi-asserted-by":"crossref","first-page":"185","DOI":"10.1121\/1.1915893","article-title":"A Scale for the Measurement of the Psychological Magnitude Pitch","volume":"8","author":"SS Stevens","year":"1937","journal-title":"J Acoust Soc Am"},{"key":"pone.0355238.ref031","unstructured":"Slaney M. \u201cAuditory toolbox,\u201d Interval Research Corporation, Tech. Rep. 1998-010. 1998."},{"key":"pone.0355238.ref032","unstructured":"Kohavi R. A study of cross-validation and bootstrap for accuracy estimation and model selection. In: Proceedings of the 14th International Joint Conference on Artificial Intelligence (IJCAI). 1995. p. 1137\u201343."},{"key":"pone.0355238.ref033","first-page":"2825","article-title":"Scikit-learn: Machine learning in Python","volume":"12","author":"F Pedregosa","year":"2011","journal-title":"J Mach Learn Res"},{"key":"pone.0355238.ref034","unstructured":"Abadi M, Barham P, Chen J, et al. Tensorflow: A system for large-scale machine learning. In: Proceedings of the 12th USENIX Symposium on Operating Systems Design and Implementation (OSDI). 2016. p. 265\u201383."},{"key":"pone.0355238.ref035","unstructured":"Chollet F, et al. Keras, software available online. 2015."},{"key":"pone.0355238.ref036","unstructured":"Streamlit. Streamlit documentation, accessed for implementation details of the demo UI. 2024."},{"issue":"9","key":"pone.0355238.ref037","doi-asserted-by":"crossref","first-page":"1263","DOI":"10.1109\/TKDE.2008.239","article-title":"Learning from Imbalanced Data","volume":"21","author":"He Haibo","year":"2009","journal-title":"IEEE Trans Knowl Data Eng"},{"key":"pone.0355238.ref038","first-page":"1","article-title":"Statistical comparisons of classifiers over multiple data sets","volume":"7","author":"J Dems\u02c7ar","year":"2006","journal-title":"J Mach Learn Res"},{"issue":"7","key":"pone.0355238.ref039","doi-asserted-by":"crossref","first-page":"1895","DOI":"10.1162\/089976698300017197","article-title":"Approximate Statistical Tests for Comparing Supervised Classification Learning Algorithms","volume":"10","author":"T Dietterich","year":"1998","journal-title":"Neural Comput"},{"issue":"1","key":"pone.0355238.ref040","doi-asserted-by":"crossref","first-page":"77","DOI":"10.1016\/S0034-4257(97)00083-7","article-title":"Selecting and interpreting measures of thematic classification accuracy","volume":"62","author":"SV Stehman","year":"1997","journal-title":"Remote Sens Environ"},{"key":"pone.0355238.ref041","unstructured":"Livingstone SR, Russo FA. The ryerson audio-visual database of emotional speech and song (ravdess). In: Proceedings of the 2018 International Conference on Affective Computing and Intelligent Interaction (ACII). 2018."},{"key":"pone.0355238.ref042","doi-asserted-by":"crossref","unstructured":"Burkhardt F, Paeschke A, Rolfes M, Sendlmeier WF, Weiss B. A database of German emotional speech. In: Interspeech 2005. 2005. p. 1517\u201320. https:\/\/doi.org\/10.21437\/interspeech.2005-446","DOI":"10.21437\/Interspeech.2005-446"},{"issue":"1","key":"pone.0355238.ref043","doi-asserted-by":"crossref","first-page":"67","DOI":"10.1109\/TAFFC.2016.2515617","article-title":"MSP-IMPROV: An Acted Corpus of Dyadic Interactions to Study Emotion Perception","volume":"8","author":"C Busso","year":"2017","journal-title":"IEEE Trans Affective Comput"},{"key":"pone.0355238.ref044","doi-asserted-by":"crossref","unstructured":"Ko T, Peddinti V, Povey D, Khudanpur S. Audio augmentation for speech recognition. In: Interspeech 2015. 2015. p. 3586\u20139. https:\/\/doi.org\/10.21437\/interspeech.2015-711","DOI":"10.21437\/Interspeech.2015-711"},{"key":"pone.0355238.ref045","unstructured":"Hendrycks D, Mu N, Cubuk ED, et al. AugMix: A simple data processing method to improve robustness and uncertainty. In: Proc. International Conference on Learning Representations (ICLR). 2020. https:\/\/doi.org\/10.48550\/arXiv.1912.02781"},{"key":"pone.0355238.ref046","unstructured":"Baevski A, Zhou H, Mohamed A, et al. \u201cwav2vec 2.0: A framework for self-supervised learning of speech representations\u201d. In: Advances in Neural Information Processing Systems (NeurIPS). 2020. https:\/\/doi.org\/10.48550\/arXiv.2006.11477"},{"key":"pone.0355238.ref047","doi-asserted-by":"crossref","first-page":"3451","DOI":"10.1109\/TASLP.2021.3122291","article-title":"HuBERT: Self-Supervised Speech Representation Learning by Masked Prediction of Hidden Units","volume":"29","author":"W-N Hsu","year":"2021","journal-title":"IEEE\/ACM Trans Audio Speech Lang Process"},{"issue":"6","key":"pone.0355238.ref048","doi-asserted-by":"crossref","first-page":"1505","DOI":"10.1109\/JSTSP.2022.3188113","article-title":"WavLM: Large-Scale Self-Supervised Pre-Training for Full Stack Speech Processing","volume":"16","author":"S Chen","year":"2022","journal-title":"IEEE J Sel Top Signal Process"},{"issue":"59","key":"pone.0355238.ref049","first-page":"1","article-title":"Domain-adversarial training of neural networks","volume":"17","author":"Y Ganin","year":"2016","journal-title":"J Mach Learn Res"},{"key":"pone.0355238.ref050","unstructured":"Guo C, Pleiss G, Sun Y, et al. On calibration of modern neural networks. In: Proceedings of the 34th International Conference on Machine Learning (ICML). 2017. p. 1321\u201330. https:\/\/doi.org\/10.48550\/arXiv.1706.04599"},{"issue":"9","key":"pone.0355238.ref051","doi-asserted-by":"crossref","first-page":"4459","DOI":"10.1007\/s00034-020-01377-y","article-title":"Excitation Features of Speech for Emotion Recognition Using Neutral Speech as Reference","volume":"39","author":"SR Kadiri","year":"2020","journal-title":"Circuits Syst Signal Process"},{"key":"pone.0355238.ref052","doi-asserted-by":"crossref","first-page":"60382","DOI":"10.1109\/ACCESS.2020.2982954","article-title":"Excitation Features of Speech for Speaker-Specific Emotion Detection","volume":"8","author":"SR Kadiri","year":"2020","journal-title":"IEEE Access"},{"key":"pone.0355238.ref053","doi-asserted-by":"crossref","unstructured":"Gr\u00f3sz T, Porjazovski D, Getman Y, Kadiri S, Kurimo M. Wav2vec2-based Paralinguistic Systems to Recognise Vocalised Emotions and Stuttering. In: Proceedings of the 30th ACM International Conference on Multimedia. 2022. p. 7026\u20139. https:\/\/doi.org\/10.1145\/3503161.3551572","DOI":"10.1145\/3503161.3551572"},{"key":"pone.0355238.ref054","doi-asserted-by":"crossref","unstructured":"Javanmardi F, Tirronen S, Kodali M, Kadiri SR, Alku P. Wav2vec-Based Detection and Severity Level Classification of Dysarthria From Speech. In: ICASSP 2023 - 2023 IEEE International Conference on Acoustics, Speech and Signal Processing (ICASSP). 2023. p. 1\u20135. https:\/\/doi.org\/10.1109\/icassp49357.2023.10094857","DOI":"10.1109\/ICASSP49357.2023.10094857"}],"container-title":["PLOS One"],"original-title":[],"language":"en","link":[{"URL":"https:\/\/dx.plos.org\/10.1371\/journal.pone.0355238","content-type":"unspecified","content-version":"vor","intended-application":"similarity-checking"}],"deposited":{"date-parts":[[2026,8,24]],"date-time":"2026-08-24T17:31:35Z","timestamp":1787592695000},"score":1,"resource":{"primary":{"URL":"https:\/\/dx.plos.org\/10.1371\/journal.pone.0355238"}},"subtitle":[],"editor":[{"given":"Sushank","family":"Chaudhary","sequence":"first","affiliation":[],"role":[{"vocabulary":"crossref","role":"editor"}]}],"short-title":[],"issued":{"date-parts":[[2026,8,24]]},"references-count":54,"journal-issue":{"issue":"8","published-online":{"date-parts":[[2026,8,24]]}},"URL":"https:\/\/doi.org\/10.1371\/journal.pone.0355238","relation":{},"ISSN":["1932-6203"],"issn-type":[{"value":"1932-6203","type":"electronic"}],"subject":[],"published":{"date-parts":[[2026,8,24]]}}}