{"id":1127,"date":"2025-04-30T07:40:12","date_gmt":"2025-04-30T07:40:12","guid":{"rendered":"https:\/\/online.binus.ac.id\/computer-science\/?p=1127"},"modified":"2025-04-30T07:40:12","modified_gmt":"2025-04-30T07:40:12","slug":"mengenali-emosi-lewat-suara-kecerdasan-buatan-kini-bisa-tebak-perasaan-anda","status":"publish","type":"post","link":"https:\/\/online.binus.ac.id\/computer-science\/2025\/04\/30\/mengenali-emosi-lewat-suara-kecerdasan-buatan-kini-bisa-tebak-perasaan-anda\/","title":{"rendered":"Mengenali Emosi Lewat Suara: Kecerdasan Buatan Kini Bisa Tebak Perasaan Anda"},"content":{"rendered":"<p class=\"\" data-start=\"269\" data-end=\"322\">Oleh: <strong data-start=\"278\" data-end=\"322\">Muhammad Fitra Kacamarga, S.Kom., M.T.I. (Faculty Member PJJ CS)<\/strong><\/p>\n<p class=\"\" data-start=\"329\" data-end=\"677\">Bayangkan jika komputer bisa tahu kapan Anda sedang senang, sedih, atau marah hanya dari suara Anda. Teknologi seperti ini bisa membuka jalan untuk interaksi manusia-komputer yang jauh lebih manusiawi. Dalam penelitian ini, kami mengembangkan sistem <strong data-start=\"579\" data-end=\"615\">Speech Emotion Recognition (SER)<\/strong>\u2014sistem cerdas yang dapat mengenali emosi dari ucapan manusia.<\/p>\n<p class=\"\" data-start=\"679\" data-end=\"823\">Dengan menggabungkan kekuatan deep learning dan perhatian khusus (attention mechanism), model kami berhasil membaca \u201crasa\u201d dari gelombang suara!<\/p>\n<p data-start=\"679\" data-end=\"823\"><img loading=\"lazy\" decoding=\"async\" class=\"alignnone size-full wp-image-1129\" src=\"https:\/\/online.binus.ac.id\/computer-science\/wp-content\/uploads\/sites\/4\/2025\/04\/Fitra2-Poster.jpg\" alt=\"\" width=\"1414\" height=\"2000\" srcset=\"https:\/\/online.binus.ac.id\/computer-science\/wp-content\/uploads\/sites\/4\/2025\/04\/Fitra2-Poster.jpg 1414w, https:\/\/online.binus.ac.id\/computer-science\/wp-content\/uploads\/sites\/4\/2025\/04\/Fitra2-Poster-212x300.jpg 212w, https:\/\/online.binus.ac.id\/computer-science\/wp-content\/uploads\/sites\/4\/2025\/04\/Fitra2-Poster-724x1024.jpg 724w, https:\/\/online.binus.ac.id\/computer-science\/wp-content\/uploads\/sites\/4\/2025\/04\/Fitra2-Poster-768x1086.jpg 768w, https:\/\/online.binus.ac.id\/computer-science\/wp-content\/uploads\/sites\/4\/2025\/04\/Fitra2-Poster-1086x1536.jpg 1086w, https:\/\/online.binus.ac.id\/computer-science\/wp-content\/uploads\/sites\/4\/2025\/04\/Fitra2-Poster-480x679.jpg 480w, https:\/\/online.binus.ac.id\/computer-science\/wp-content\/uploads\/sites\/4\/2025\/04\/Fitra2-Poster-1024x1448.jpg 1024w\" sizes=\"auto, (max-width: 1414px) 100vw, 1414px\" \/><\/p>\n<h3 class=\"\" data-start=\"830\" data-end=\"877\">\ud83e\udde0 Model Canggih: CRNN + BI-GRU + Attention<\/h3>\n<p class=\"\" data-start=\"879\" data-end=\"931\">Kami menggunakan kombinasi arsitektur deep learning:<\/p>\n<ul data-start=\"933\" data-end=\"1196\">\n<li class=\"\" data-start=\"933\" data-end=\"1012\">\n<p class=\"\" data-start=\"935\" data-end=\"1012\"><strong data-start=\"935\" data-end=\"984\">Convolutional Recurrent Neural Network (CRNN)<\/strong> untuk menangkap pola suara.<\/p>\n<\/li>\n<li class=\"\" data-start=\"1013\" data-end=\"1100\">\n<p class=\"\" data-start=\"1015\" data-end=\"1100\"><strong data-start=\"1015\" data-end=\"1062\">Bidirectional Gated Recurrent Unit (Bi-GRU)<\/strong> untuk memahami konteks dari dua arah.<\/p>\n<\/li>\n<li class=\"\" data-start=\"1101\" data-end=\"1196\">\n<p class=\"\" data-start=\"1103\" data-end=\"1196\"><strong data-start=\"1103\" data-end=\"1135\">Bahdanau Attention Mechanism<\/strong> untuk memberi \u201cfokus\u201d pada bagian penting dari sinyal suara.<\/p>\n<\/li>\n<\/ul>\n<p class=\"\" data-start=\"1198\" data-end=\"1369\">Sinyal suara diolah dan ditransformasikan menjadi fitur-fitur yang kemudian dianalisis untuk menghasilkan prediksi emosi\u2014apakah pembicara sedang marah, senang, takut, dsb.<\/p>\n<h3 class=\"\" data-start=\"1376\" data-end=\"1403\">\ud83d\udd0a Apa Saja yang Diuji?<\/h3>\n<p class=\"\" data-start=\"1405\" data-end=\"1600\">Model diuji menggunakan dataset <strong data-start=\"1437\" data-end=\"1448\">RAVDESS<\/strong> yang terdiri dari rekaman suara dengan berbagai ekspresi emosional. Untuk memperkuat model, kami juga menggunakan teknik <strong data-start=\"1570\" data-end=\"1591\">data augmentation<\/strong> seperti:<\/p>\n<ul data-start=\"1602\" data-end=\"1673\">\n<li class=\"\" data-start=\"1602\" data-end=\"1635\">\n<p class=\"\" data-start=\"1604\" data-end=\"1635\">Penambahan noise latar belakang<\/p>\n<\/li>\n<li class=\"\" data-start=\"1636\" data-end=\"1673\">\n<p class=\"\" data-start=\"1638\" data-end=\"1673\">Perubahan kecepatan dan pitch suara<\/p>\n<\/li>\n<\/ul>\n<h3 class=\"\" data-start=\"1680\" data-end=\"1709\">\ud83d\udcc8 Hasil yang Mengesankan<\/h3>\n<p class=\"\" data-start=\"1711\" data-end=\"1787\">Model kami mencetak akurasi tertinggi dibandingkan metode-metode sebelumnya:<\/p>\n<ul data-start=\"1789\" data-end=\"1864\">\n<li class=\"\" data-start=\"1789\" data-end=\"1827\">\n<p class=\"\" data-start=\"1791\" data-end=\"1827\"><strong data-start=\"1791\" data-end=\"1819\">Unweighted Accuracy (UA)<\/strong>: 90,19%<\/p>\n<\/li>\n<li class=\"\" data-start=\"1828\" data-end=\"1864\">\n<p class=\"\" data-start=\"1830\" data-end=\"1864\"><strong data-start=\"1830\" data-end=\"1856\">Weighted Accuracy (WA)<\/strong>: 90,53%<\/p>\n<\/li>\n<\/ul>\n<p class=\"\" data-start=\"1866\" data-end=\"1975\">Artinya, model mampu mengenali emosi lebih akurat dan konsisten meski data berasal dari berbagai jenis suara.<\/p>\n<h3 class=\"\" data-start=\"1982\" data-end=\"2003\">\ud83d\udd0d Temuan Menarik<\/h3>\n<ul data-start=\"2005\" data-end=\"2293\">\n<li class=\"\" data-start=\"2005\" data-end=\"2079\">\n<p class=\"\" data-start=\"2007\" data-end=\"2079\">Model kami unggul hingga <strong data-start=\"2032\" data-end=\"2053\">11\u201312% lebih baik<\/strong> dibandingkan metode lain.<\/p>\n<\/li>\n<li class=\"\" data-start=\"2080\" data-end=\"2187\">\n<p class=\"\" data-start=\"2082\" data-end=\"2187\">Pendekatan <strong data-start=\"2093\" data-end=\"2114\">few-shot learning<\/strong> terbukti meningkatkan kinerja, terutama saat data emosi sangat terbatas.<\/p>\n<\/li>\n<li class=\"\" data-start=\"2188\" data-end=\"2293\">\n<p class=\"\" data-start=\"2190\" data-end=\"2293\">Text-based emotion recognition (berbasis teks) masih kesulitan menandingi akurasi model berbasis suara.<\/p>\n<\/li>\n<\/ul>\n<h3 class=\"\" data-start=\"2300\" data-end=\"2317\">\ud83d\udccc Kesimpulan<\/h3>\n<p class=\"\" data-start=\"2319\" data-end=\"2539\">Penelitian ini membuka jalan untuk masa depan interaksi cerdas antara manusia dan mesin. Dari asisten virtual yang lebih empatik, hingga sistem call center yang memahami emosi pelanggan\u2014teknologi ini punya potensi besar.<\/p>\n<p class=\"\" data-start=\"2546\" data-end=\"2627\">\ud83d\udca1 <em data-start=\"2549\" data-end=\"2627\">Masa depan bukan sekadar memahami kata-kata, tapi juga perasaan di baliknya.<\/em><\/p>\n<div>\n<p>Sumber: Penelitian Dosen dan Mahasiswa PJJ CS<\/p>\n<p>Editor: Pandu Dwi Luhur Pambudi, S.Kom., M.Kom., M.I.M<\/p>\n<\/div>\n<h3>#BINUSRESEARCHPOINT #TEKNIKINFORMATIKA #COMPUTERSCIENCE #BINUS #BINUSUNIVERSITY<\/h3>\n","protected":false},"excerpt":{"rendered":"<p>Oleh: Muhammad Fitra Kacamarga, S.Kom., M.T.I. (Faculty Member PJJ CS) Bayangkan jika komputer bisa tahu kapan Anda sedang senang, sedih, atau marah hanya dari suara Anda. Teknologi seperti ini bisa membuka jalan untuk interaksi manusia-komputer yang jauh lebih manusiawi. Dalam penelitian ini, kami mengembangkan sistem Speech Emotion Recognition (SER)\u2014sistem cerdas yang dapat mengenali emosi dari [&hellip;]<\/p>\n","protected":false},"author":702,"featured_media":1129,"comment_status":"open","ping_status":"open","sticky":false,"template":"","format":"standard","meta":{"footnotes":""},"categories":[2],"tags":[16],"class_list":["post-1127","post","type-post","status-publish","format-standard","has-post-thumbnail","hentry","category-article","tag-binusresearchpoint-teknikinformatika-computerscience-binus-binusuniversity"],"_links":{"self":[{"href":"https:\/\/online.binus.ac.id\/computer-science\/wp-json\/wp\/v2\/posts\/1127","targetHints":{"allow":["GET"]}}],"collection":[{"href":"https:\/\/online.binus.ac.id\/computer-science\/wp-json\/wp\/v2\/posts"}],"about":[{"href":"https:\/\/online.binus.ac.id\/computer-science\/wp-json\/wp\/v2\/types\/post"}],"author":[{"embeddable":true,"href":"https:\/\/online.binus.ac.id\/computer-science\/wp-json\/wp\/v2\/users\/702"}],"replies":[{"embeddable":true,"href":"https:\/\/online.binus.ac.id\/computer-science\/wp-json\/wp\/v2\/comments?post=1127"}],"version-history":[{"count":1,"href":"https:\/\/online.binus.ac.id\/computer-science\/wp-json\/wp\/v2\/posts\/1127\/revisions"}],"predecessor-version":[{"id":1134,"href":"https:\/\/online.binus.ac.id\/computer-science\/wp-json\/wp\/v2\/posts\/1127\/revisions\/1134"}],"wp:featuredmedia":[{"embeddable":true,"href":"https:\/\/online.binus.ac.id\/computer-science\/wp-json\/wp\/v2\/media\/1129"}],"wp:attachment":[{"href":"https:\/\/online.binus.ac.id\/computer-science\/wp-json\/wp\/v2\/media?parent=1127"}],"wp:term":[{"taxonomy":"category","embeddable":true,"href":"https:\/\/online.binus.ac.id\/computer-science\/wp-json\/wp\/v2\/categories?post=1127"},{"taxonomy":"post_tag","embeddable":true,"href":"https:\/\/online.binus.ac.id\/computer-science\/wp-json\/wp\/v2\/tags?post=1127"}],"curies":[{"name":"wp","href":"https:\/\/api.w.org\/{rel}","templated":true}]}}