{"status":"ok","message-type":"work","message-version":"1.0.0","message":{"indexed":{"date-parts":[[2026,7,16]],"date-time":"2026-07-16T10:07:17Z","timestamp":1784196437420,"version":"3.55.0"},"reference-count":45,"publisher":"Wiley","issue":"8","license":[{"start":{"date-parts":[[2026,6,18]],"date-time":"2026-06-18T00:00:00Z","timestamp":1781740800000},"content-version":"vor","delay-in-days":0,"URL":"http:\/\/onlinelibrary.wiley.com\/termsAndConditions#vor"},{"start":{"date-parts":[[2026,6,18]],"date-time":"2026-06-18T00:00:00Z","timestamp":1781740800000},"content-version":"tdm","delay-in-days":0,"URL":"http:\/\/doi.wiley.com\/10.1002\/tdm_license_1.1"}],"content-domain":{"domain":["onlinelibrary.wiley.com"],"crossmark-restriction":true},"short-container-title":["Expert Systems"],"published-print":{"date-parts":[[2026,8]]},"abstract":"<jats:title>ABSTRACT<\/jats:title>\n                  <jats:p>The rise of Emotion AI is transforming how human emotions are detected, interpreted and responded to in real\u2010time social systems. Speech emotion recognition (SER) enables machines to understand human emotions from voice, fostering more empathetic and context\u2010aware interactions. However, the currently deployed SER systems are frequently based on highly resource\u2010consuming large models, which are unsuitable for real\u2010time deployment on edge devices with limited memory and processing power. Most previous studies have focused on single\u2010language or single accent datasets, which results in ineffective extrapolation to various speakers, accents and the real world. This study introduces a scalable and lightweight SER system designed for the TinyML environment and suitable for deployment on resource\u2010constrained systems, including social networks using IoT technologies, assistive technologies and embedded mental health devices. Combining RAVDESS, TESS and SAVEE increases dataset diversity. Their effectiveness in capturing both spectral and temporal emotion cues is tested across six hybrid frameworks of deep learning, such as CNN\u2009+\u2009BiLSTM and CNN\u2009+\u2009BiGRU with multi\u2010head attention. The best model achieved 74.15% accuracy, with a macro F1\u2010score of 0.73, a weighted F1\u2010score of 0.74, and a highest class\u2010level F1\u2010score of 0.82, supporting low\u2010latency emotion recognition for affect\u2010aware edge applications. The quantization through TensorFlow Lite further reduces the model size by up to 94.5% and achieves inference latency as low as 3.4\u2009ms, making it suitable for deployment on microcontrollers. This study contributes to Emotion AI as it allows detecting emotions on edge devices to facilitate affect\u2010aware customer service, support mental health, improve education and more broadly, computational social systems.<\/jats:p>","DOI":"10.1111\/exsy.70322","type":"journal-article","created":{"date-parts":[[2026,6,19]],"date-time":"2026-06-19T00:32:17Z","timestamp":1781829137000},"update-policy":"https:\/\/doi.org\/10.1002\/crossmark_policy","source":"Crossref","is-referenced-by-count":0,"title":["Emotion\n                    <scp>AI<\/scp>\n                    on the Edge: A\n                    <scp>TinyML<\/scp>\n                    \u2010Driven Framework for Speech Emotion Recognition in Social Environments"],"prefix":"10.1111","volume":"43","author":[{"given":"Md. Sakib Bin","family":"Alam","sequence":"first","affiliation":[{"name":"School of Information and Communication Technology Griffith University  Gold Coast Queensland Australia"}],"role":[{"vocabulary":"crossref","role":"author"}]},{"given":"Aiman","family":"Lameesa","sequence":"additional","affiliation":[{"name":"School of Information and Communication Technology Griffith University  Gold Coast Queensland Australia"}],"role":[{"vocabulary":"crossref","role":"author"}]},{"given":"Barsha","family":"Roy","sequence":"additional","affiliation":[{"name":"Department of Computer Science University of Central Florida  Orlando Florida USA"}],"role":[{"vocabulary":"crossref","role":"author"}]},{"ORCID":"https:\/\/orcid.org\/0000-0003-3360-3636","authenticated-orcid":false,"given":"Shams Forruque","family":"Ahmed","sequence":"additional","affiliation":[{"name":"School of Mathematical Sciences Sunway University  Bandar Sunway, Petaling Jaya Selangor Darul Ehsan Malaysia"},{"name":"Miyan Research Institute International University of Business Agriculture and Technology  Dhaka Bangladesh"}],"role":[{"vocabulary":"crossref","role":"author"}]},{"ORCID":"https:\/\/orcid.org\/0000-0002-2798-0104","authenticated-orcid":false,"given":"Amir H.","family":"Gandomi","sequence":"additional","affiliation":[{"name":"Faculty of Engineering &amp; Information Technology University of Technology Sydney  Sydney New South Wales Australia"},{"name":"University Research and Innovation Center (EKIK), \u00d3buda University  Budapest Hungary"}],"role":[{"vocabulary":"crossref","role":"author"}]}],"member":"311","published-online":{"date-parts":[[2026,6,18]]},"reference":[{"key":"e_1_2_9_2_1","doi-asserted-by":"publisher","DOI":"10.1109\/ACCESS.2023.3294111"},{"key":"e_1_2_9_3_1","doi-asserted-by":"publisher","DOI":"10.1109\/ACCESS.2024.3422480"},{"key":"e_1_2_9_4_1","doi-asserted-by":"publisher","DOI":"10.1007\/s10462\u2010023\u201010466\u20108"},{"key":"e_1_2_9_5_1","doi-asserted-by":"publisher","DOI":"10.1038\/s41598\u2010024\u201063776\u20104"},{"key":"e_1_2_9_6_1","doi-asserted-by":"publisher","DOI":"10.3390\/app13084750"},{"key":"e_1_2_9_7_1","doi-asserted-by":"publisher","DOI":"10.1016\/j.eswa.2022.119016"},{"key":"e_1_2_9_8_1","doi-asserted-by":"publisher","DOI":"10.1016\/j.knosys.2023.111077"},{"key":"e_1_2_9_9_1","doi-asserted-by":"publisher","DOI":"10.1016\/j.iot.2021.100461"},{"key":"e_1_2_9_10_1","doi-asserted-by":"publisher","DOI":"10.1016\/j.neucom.2024.128879"},{"key":"e_1_2_9_11_1","doi-asserted-by":"publisher","DOI":"10.1109\/ACIIW59127.2023.10388120"},{"key":"e_1_2_9_12_1","doi-asserted-by":"publisher","DOI":"10.1016\/j.neucom.2023.127015"},{"key":"e_1_2_9_13_1","unstructured":"Ghamari S. K.Ozcan T.Dinh et\u00a0al.2021.\u201cQuantization\u2010Guided Training for Compact TinyML Models.\u201dhttp:\/\/arxiv.org\/abs\/2103.06231."},{"key":"e_1_2_9_14_1","doi-asserted-by":"publisher","DOI":"10.1016\/j.specom.2025.103201"},{"key":"e_1_2_9_15_1","unstructured":"Haq S. andP. J. B.Jackson.2009.\u201cSpeaker\u2010Dependent Audio\u2010Visual Emotion Recognition.\u201dAuditory\u2010Visual Speech Processing AVSP 2009 2009."},{"key":"e_1_2_9_16_1","doi-asserted-by":"publisher","DOI":"10.1016\/j.ijcce.2024.11.008"},{"key":"e_1_2_9_17_1","doi-asserted-by":"publisher","DOI":"10.1016\/j.eswa.2023.122946"},{"key":"e_1_2_9_18_1","doi-asserted-by":"publisher","DOI":"10.1016\/j.prime.2024.100707"},{"key":"e_1_2_9_19_1","doi-asserted-by":"publisher","DOI":"10.1016\/j.specom.2024.103184"},{"key":"e_1_2_9_20_1","doi-asserted-by":"publisher","DOI":"10.1016\/j.specom.2023.103010"},{"key":"e_1_2_9_21_1","doi-asserted-by":"publisher","DOI":"10.1016\/j.neucom.2024.128711"},{"key":"e_1_2_9_22_1","doi-asserted-by":"publisher","DOI":"10.1109\/TASLP.2023.3282092"},{"key":"e_1_2_9_23_1","doi-asserted-by":"publisher","DOI":"10.1016\/j.apacoust.2022.109178"},{"key":"e_1_2_9_24_1","doi-asserted-by":"publisher","DOI":"10.1371\/journal.pone.0196391"},{"key":"e_1_2_9_25_1","doi-asserted-by":"publisher","DOI":"10.1145\/3681716.3681725"},{"key":"e_1_2_9_26_1","doi-asserted-by":"publisher","DOI":"10.1007\/s44163\u2010023\u201000051\u2010x"},{"key":"e_1_2_9_27_1","doi-asserted-by":"publisher","DOI":"10.1016\/j.iswa.2023.200266"},{"key":"e_1_2_9_28_1","doi-asserted-by":"publisher","DOI":"10.1016\/j.bspc.2024.106201"},{"key":"e_1_2_9_29_1","doi-asserted-by":"publisher","DOI":"10.1007\/s11042\u2010023\u201016849\u2010x"},{"key":"e_1_2_9_30_1","doi-asserted-by":"publisher","DOI":"10.1007\/s11760\u2010023\u201002716\u20107"},{"key":"e_1_2_9_31_1","doi-asserted-by":"publisher","DOI":"10.1109\/TAFFC.2025.3537991"},{"key":"e_1_2_9_32_1","doi-asserted-by":"publisher","DOI":"10.1007\/s11760\u2010024\u201003574\u20107"},{"key":"e_1_2_9_33_1","doi-asserted-by":"publisher","DOI":"10.1038\/s41598\u2010024\u201052989\u20102"},{"key":"e_1_2_9_34_1","unstructured":"Pichora\u2010Fuller M. K. andK.Dupuis.2020.\u201cToronto Emotional Speech Set (TESS).\u201dScholars Portal Dataverse 1."},{"key":"e_1_2_9_35_1","doi-asserted-by":"publisher","DOI":"10.1016\/j.neucom.2024.128646"},{"key":"e_1_2_9_36_1","doi-asserted-by":"publisher","DOI":"10.1016\/j.bspc.2025.107636"},{"key":"e_1_2_9_37_1","doi-asserted-by":"publisher","DOI":"10.1016\/j.specom.2024.103102"},{"key":"e_1_2_9_38_1","doi-asserted-by":"publisher","DOI":"10.1016\/j.eswa.2023.119633"},{"key":"e_1_2_9_39_1","doi-asserted-by":"publisher","DOI":"10.1016\/j.asoc.2024.111735"},{"key":"e_1_2_9_40_1","doi-asserted-by":"publisher","DOI":"10.1016\/j.iswa.2024.200351"},{"key":"e_1_2_9_41_1","doi-asserted-by":"publisher","DOI":"10.1016\/j.bspc.2025.107860"},{"key":"e_1_2_9_42_1","doi-asserted-by":"publisher","DOI":"10.1016\/j.apacoust.2024.109886"},{"key":"e_1_2_9_43_1","doi-asserted-by":"publisher","DOI":"10.1371\/journal.pone.0318297"},{"key":"e_1_2_9_44_1","doi-asserted-by":"publisher","DOI":"10.1016\/j.neucom.2024.128177"},{"key":"e_1_2_9_45_1","doi-asserted-by":"publisher","DOI":"10.1016\/j.engappai.2025.110060"},{"key":"e_1_2_9_46_1","doi-asserted-by":"publisher","DOI":"10.1016\/j.bspc.2025.107687"}],"container-title":["Expert Systems"],"original-title":[],"language":"en","link":[{"URL":"https:\/\/onlinelibrary.wiley.com\/doi\/pdf\/10.1111\/exsy.70322","content-type":"application\/pdf","content-version":"vor","intended-application":"text-mining"},{"URL":"https:\/\/onlinelibrary.wiley.com\/doi\/full-xml\/10.1111\/exsy.70322","content-type":"application\/xml","content-version":"vor","intended-application":"text-mining"},{"URL":"https:\/\/onlinelibrary.wiley.com\/doi\/pdf\/10.1111\/exsy.70322","content-type":"unspecified","content-version":"vor","intended-application":"similarity-checking"}],"deposited":{"date-parts":[[2026,7,16]],"date-time":"2026-07-16T09:34:33Z","timestamp":1784194473000},"score":1,"resource":{"primary":{"URL":"https:\/\/onlinelibrary.wiley.com\/doi\/10.1111\/exsy.70322"}},"subtitle":[],"short-title":[],"issued":{"date-parts":[[2026,6,18]]},"references-count":45,"journal-issue":{"issue":"8","published-print":{"date-parts":[[2026,8]]}},"alternative-id":["10.1111\/exsy.70322"],"URL":"https:\/\/doi.org\/10.1111\/exsy.70322","archive":["Portico"],"relation":{},"ISSN":["0266-4720","1468-0394"],"issn-type":[{"value":"0266-4720","type":"print"},{"value":"1468-0394","type":"electronic"}],"subject":[],"published":{"date-parts":[[2026,6,18]]},"assertion":[{"value":"2025-08-30","order":0,"name":"received","label":"Received","group":{"name":"publication_history","label":"Publication History"}},{"value":"2026-05-23","order":2,"name":"accepted","label":"Accepted","group":{"name":"publication_history","label":"Publication History"}},{"value":"2026-06-18","order":3,"name":"published","label":"Published","group":{"name":"publication_history","label":"Publication History"}}],"article-number":"e70322"}}