SuperKogito/SER-datasets：一组用于语音中情感识别/检测的数据集

项目介绍

SuperKogito/SER-datasets 是一个托管在 GitHub 上的开源项目，主要使用 HTML 编写，采用 MIT License 许可证，目前获得 427 个 Star、49 次 Fork。

仓库全名: SuperKogito/SER-datasets

README 摘要

Speech Emotion Recognition (SER) Datasets: A collection of datasets (count=57) for the purpose of emotion recognition/detection in speech. The table is chronologically ordered and includes a description of the content of each dataset along with the emotions included. The table can be browsed, sorted and searched under https://superkogito.github.io/SER-datasets/ Dataset | Year | Content | Emotions | Format | Size |…

完整内容见 GitHub 原始 README 或本页。

常见问题

SER-datasets 用什么语言开发？

主要使用 HTML 编写，采用 MIT License 许可证，目前有 427 个 Star。

SER-datasets 可以免费商用吗？

如何获取 SER-datasets 的源码？

可以直接克隆仓库：git clone https://github.com/SuperKogito/SER-datasets.git，默认分支为 master；安装与运行步骤见仓库 README。

SER-datasets 最近还在维护吗？

仓库最近一次更新为 2024年10月1日，当前有 18 个未关闭 Issue；Star 数为 427，可作为社区活跃度的参考。

关于

GitHub 原始描述（英文）

A collection of datasets for the purpose of emotion recognition/detection in speech.

audio audio-datasets datasets emotions emotions-recognition multimodal-emotion-recognition speech speech-emotion-recognition

MIT License
427 Stars
49 Forks
14 Watching

描述取自 GitHub 官方 API，保持仓库作者书写的原文。

简要分析

热度分

539

价值分

271

活跃状态

活跃

GitHub

语言占比

Batchfile

HTML

JavaScript

Makefile

Python

Shell

Release

v4.0.02024年6月24日
v3.3.02024年6月9日
v3.2.02024年6月4日
v3.1.02023年3月15日
v3.0.02023年2月14日

贡献者

Speech Emotion Recognition (SER) Datasets: A collection of datasets (count=57) for the purpose of emotion recognition/detection in speech. The table is chronologically ordered and includes a description of the content of each dataset along with the emotions included. The table can be browsed, sorted and searched under https://superkogito.github.io/SER-datasets/

Dataset	Year	Content	Emotions	Format	Size	Language	Paper	Access	License
_{Hi, KIA}	₂₀₂₂	_{A shared short Wakeup Word database focusing on perceived emotion in speech The dataset contains 488 Wakeup Word speech}	_{angry, happy, sad, neutral}	_Audio	_{0.75 GB}	_Korean	_{Hi, KIA: A Speech Emotion Recognition Dataset for Wake-Up Words}	_Open	_{CC BY-SA 4.0}
_{Emozionalmente}	₂₀₂₂	_{6902 labeled samples acted out by 431 amateur actors while verbalizing 18 different sentences}	_{anger, disgust, fear, joy, sadness, surprise, neutral}	_Audio	_{0.581 GB}	_Italian	_–	_Open	_{CC BY 4.0}
_Kannada	₂₀₂₂	_{468 audio samples, six different sentences, pronounced by thirteen people (four male and nine female), in five basic emotions plus one neutral emotion}	_{Anger, Sadness, Surprise, Happiness, Fear, Neutral}	_Audio	_{0.1661 GB}	_Kannada	_–	_Open	_{CC BY 4.0}
_emoUERJ	₂₀₂₁	_{Ten sentences from eight actors, equally divided between genders, and they were free to choose the phrases for record audios in four emotions (377 audios).}	_{happiness, anger, sadness or neutral}	_Audio	_{0.1051 GB}	_{Portuguese (Brazilian)}	_–	_Open	_{CC BY 4.0}
_ESCorpus-PE	₂₀₂₁	_{Spanish peruvian speech gathered from Spanish interviews, TV reports, political debate and testimonials. It contains 3749 utterances, 80 speakers (44 male and 36 female), created from Youtube audios}	_{Valence, Arousal and Dominance}	_Audio	_{1.9 GB}	_{Spanish (Peruvian)}	_–	_Open	_{CC BY-SA 4.0}
_SUBSECO	₂₀₂₁	_{7000 sentence-level utterances of the Bangla language, 20 professional actors (10 males and 10 females), recordings, 10 sentences for 7 target emotions.}	_{Anger, Disgust, Fear, Happiness, Neutral, Sadness and Surprise}	_Audio	_{1.7 GB}	_English	_{SUST Bangla Emotional Speech Corpus (SUBESCO): An audio-only emotional speech corpus for Bangla}	_Open	_{CC BY 4.0}
_{Audio-Speech-Sentiment}	₂₀₂₁	_{Audio Speech Sentiment Dataset}	_{4 emotions provides audio recordings of spoken sentences for anger, happiness, sadness, and neutral emotions.}	_Audio	_{1.1 GB}	_English	_–	_Open	_{CC0: Public Domain}
_Quechua-SER	₂₀₂₂	_{12420 audio recordings (~15 hours) and their transcriptions by 7 native speakers.}	_{Emotional labels using dimensions: valence, arousal, and dominance.}	_Audio	_{3.53 GB}	_{Quechua Collao}	_{A speech corpus of Quechua Collao for automatic dimensional emotion recognition}	_Open	_{CC BY 4.0}
_MESD	₂₀₂₂	_{864 audio files of single-word emotional utterances with Mexican cultural shaping.}	_{6 emotions provides single-word utterances for anger, disgust, fear, happiness, neutral, and sadness.}	_Audio	_{0.097 GB}	_{Spanish (Mexican)}	_{The Mexican Emotional Speech Database (MESD): elaboration and assessment based on machine learning}	_Open	_{CC BY 4.0}
_SyntAct	₂₀₂₂	_{Synthesized database of three basic emotions and neutral expression based on rule-based manipulation for a diphone synthesizer which we release to the public}	_{997 utterances including 6 emotions: angry, bored, happy, neutral, sad and scared}	_Audio	_{0.941 GB}	_German	_{SyntAct: A Synthesized Database of Basic Emotions}	_Open	_{CC BY-SA 4.0}
_LSSED	₂₀₂₁	_{LSSED: A Large-Scale Dataset and Benchmark for Speech Emotion Recognition}	_{Anger, happiness, sadness, disappointment, boredom, disgust, excitement, fear, surprise, normal, and other.}	_Audio	_{90 GB}	_English	_{LSSED: A Large-Scale Spanish Emotional Speech Database for Speech Processing and Machine Learning}	_Restricted	_-
_MLEnd	₂₀₂₁	_{~32700 audio recordings files produced by 154 speakers. Each audio recording corresponds to one English numeral (from “zero” to “billion”)}	_{Intonations: neutral, bored, excited and question}	_Audio	_{2.27 GB}	_–	_–	_Open	_Unknown
_ASVP-ESD	₂₀₂₁	_{~13285 audio files collected from movies, tv shows and youtube containing speech and non-speech.}	_{12 different natural emotions (boredom, neutral, happiness, sadness, anger, fear, surprise, disgust, excitement, pleasure, pain, disappointment) with 2 levels of intensity.}	_Audio	_{2 GB}	_{Chinese, English, French, Russian and others}	_–	_Open	_Unknown
_ESD	₂₀₂₁	_{29 hours, 3500 sentences, by 10 native English speakers and 10 native Chinese speakers.}	_{5 emotions: angry, happy, neutral, sad, and surprise.}	_{Audio, Text}	_{2.4 GB}	_{Chinese, English}	_{Seen And Unseen Emotional Style Transfer For Voice Conversion With A New Emotional Speech Dataset}	_Open	_{Academic License}
_MuSe-CAR	₂₀₂₁	_{40 hours, 6,000+ recordings of 25,000+ sentences by 70+ English speakers (see db link for details).}	_{continuous emotion dimensions characterized using valence, arousal, and trustworthiness.}	_{Audio, Video, Text}	_{15 GB}	_English	_{The Multimodal Sentiment Analysis in Car Reviews (MuSe-CaR) Dataset: Collection, Insights and Improvements}	_Restricted	_{Academic License & Commercial License}
_{THAI SER}	₂₀₂₁	_{The recordings are 41 hours, 36 minutes long (27,854 utterances), and were performed by 200 professional actors (112 female, 88 male).}	_{5 main emotions assigned to actors: Neutral, Anger, Happiness, Sadness, and Frustration.}	_Audio	_{12 GB}	_Thai	_–	_Open	_{CC BY-SA 4.0}
_{French Emotional Speech Database - Oréau}	₂₀₂₀	_{79 utterances with 10 to 13 utterances pro emotion by 32 non-professional speakers.}	_{7 emotions: sadness, anger, disgust, fear, surprise, joy, neutral.}	_Audio	_{0.264 GB}	_French	_–	_Open	_{CC BY 4.0}
_Att-HACK	₂₀₂₀	_{25 speakers interpreting 100 utterances in 4 social attitudes, with 3-5 repetitions each per attitude for a total of around 30 hours of speech.}	_{expressive speech in French, 100 phrases with multiple versions (3 to 5) in four social attitudes (friendly, distant, dominant and seductive).}	_Audio	_{6.6 GB}	_French	_{Att-HACK: An Expressive Speech Database with Social Attitudes}	_Open	_{CC BY-NC-ND 4.0}
_{MSP-Podcast corpus}	₂₀₂₀	_{100 hours by over 100 speakers (see db link for details).}	_{This corpus is annotated with emotional labels using attribute-based descriptors (activation, dominance and valence) and categorical labels (anger, happiness, sadness, disgust, surprised, fear, contempt, neutral and other).}	_Audio	_{13.4 GB}	_English	_{The MSP-Conversation Corpus}	_Restricted	_{Academic License & Commercial License}
_BEASC	₂₀₂₀	_{Bangla Emotional Audio-Speech Corpus}	_{6 emotions provides Bangla spoken utterances for anger, happiness, sadness, fear, surprise, and neutral.}	_Audio	_{9 GB}	_Bangla	_{BEASC: Bangla Emotional Audio-Speech Corpus - A Speech Emotion Recognition Corpus for the Low-Resource Bangla Language}	_Open	_{CC BY 4.0}
_{emotiontts open db}	₂₀₂₀	_{Recordings and their associated transcriptions by a diverse group of speakers.}	_{4 emotions: general, joy, anger, and sadness.}	_{Audio, Text}	_–	_Korean	_–	_{Partially open}	_{CC BY-NC-SA 4.0}
_URDU-Dataset	₂₀₂₀	_{400 utterances by 38 speakers (27 male and 11 female).}	_{4 emotions: angry, happy, neutral, and sad.}	_Audio	_{0.072 GB}	_Urdu	_{Cross Lingual Speech Emotion Recognition: Urdu vs. Western Languages}	_Open	_–
_BAVED	₂₀₂₀	_{1935 recording by 61 speakers (45 male and 16 female).}	_{3 levels of emotion.}	_Audio	_{0.195 GB}	_Arabic	_–	_Open	_–
_VIVAE	₂₀₂₀	_{non-speech, 1085 audio file by 11 speakers.}	_{non-speech 6 emotions: achievement, anger, fear, pain, pleasure, and surprise with 3 emotional intensities (low, moderate, strong, peak).}	_Audio	_{0.0935 GB}	_{Nonverbal (English)}	_{The Variably Intense Vocalizations of Affect and Emotion (VIVAE) corpus prompts new perspective on nonspeech perception}	_Restricted	_{CC BY-NC-SA 4.0}
_PMEmo	₂₀₁₉	_{Dataset containing emotion annotations of 794 songs as well as the simultaneous electrodermal activity (EDA) signals. A Music Emotion Experiment was well-designed for collecting the affective-annotated music corpus of high quality, which recruited 457 subjects.}	_{Valence, Arousal}	_{Audio, EDA}	_{1.3 GB}	_{Chinese, English}	_{The PMEmo Dataset for Music Emotion Recognition}	_Open	_{CC BY-SA 4.0}
_SEWA	₂₀₁₉	_{more than 2000 minutes of audio-visual data of 398 people (201 male and 197 female) coming from 6 cultures.}	_{emotions are characterized using valence and arousal.}	_{Audio, Video}	_–	_{Chinese, English, German, Greek, Hungarian and Serbian}	_{SEWA DB: A Rich Database for Audio-Visual Emotion and Sentiment Research in the Wild}	_Restricted	_{SEWA EULA}
_MELD	₂₀₁₉	_{1400 dialogues and 14000 utterances from Friends TV series by multiple speakers.}	_{7 emotions: Anger, disgust, sadness, joy, neutral, surprise and fear. MELD also has sentiment (positive, negative and neutral) annotation for each utterance.}	_{Audio, Video, Text}	_{10.1 GB}	_English	_{MELD: A Multimodal Multi-Party Dataset for Emotion Recognition in Conversations}	_Open	_{MELD: GPL-3.0 License}
_ShEMO	₂₀₁₉	_{3000 semi-natural utterances, equivalent to 3 hours and 25 minutes of speech data from online radio plays by 87 native-Persian speakers.}	_{6 emotions: anger, fear, happiness, sadness, neutral and surprise.}	_Audio	_{0.101 GB}	_Persian	_{ShEMO: a large-scale validated database for Persian speech emotion detection}	_Open	_–
_DEMoS	₂₀₁₉	_{9365 emotional and 332 neutral samples produced by 68 native speakers (23 females, 45 males).}	_{7/6 emotions: anger, sadness, happiness, fear, surprise, disgust, and the secondary emotion guilt.}	_Audio	_{2.5 GB}	_Italian	_{DEMoS: An Italian emotional speech corpus. Elicitation methods, machine learning, and perception}	_Restricted	_{EULA: End User License Agreement}
_AESDD	₂₀₁₈	_{around 500 utterances by a diverse group of actors (over 5 actors) siumlating various emotions.}	_{5 emotions: anger, disgust, fear, happiness, and sadness.}	_Audio	_{0.392 GB}	_Greek	_{Speech Emotion Recognition for Performance Interaction}	_Open	_–
_Emov-DB	₂₀₁₈	_{Recordings for 4 speakers- 2 males and 2 females.}	_{The emotional styles are neutral, sleepiness, anger, disgust and amused.}	_Audio	_{5.88 GB}	_English	_{The emotional voices database: Towards controlling the emotion dimension in voice generation systems}	_Open	_–
_RAVDESS	₂₀₁₈	_{7356 recordings by 24 actors.}	_{7 emotions: calm, happy, sad, angry, fearful, surprise, and disgust}	_{Audio, Video}	_{24.8 GB}	_English	_{The Ryerson Audio-Visual Database of Emotional Speech and Song (RAVDESS): A dynamic, multimodal set of facial and vocal expressions in North American English}	_Open	_{CC BY-NC-SA 4.0}
_{JL corpus}	₂₀₁₈	_{2400 recording of 240 sentences by 4 actors (2 males and 2 females).}	_{5 primary emotions: angry, sad, neutral, happy, excited. 5 secondary emotions: anxious, apologetic, pensive, worried, enthusiastic.}	_Audio	_{1.9 GB}	_English	_{An Open Source Emotional Speech Corpus for Human Robot Interaction Applications}	_Open	_{CC0 1.0}
_CaFE	₂₀₁₈	_{6 different sentences by 12 speakers (6 fmelaes + 6 males).}	_{7 emotions: happy, sad, angry, fearful, surprise, disgust and neutral. Each emotion is acted in 2 different intensities.}	_Audio	_{2 GB}	_{French (Canadian)}	_–	_Open	_{CC BY-NC-SA 4.0}
_EmoFilm	₂₀₁₈	_{1115 audio instances sentences extracted from various films.}	_{5 emotions: anger, contempt, happiness, fear, and sadness.}	_Audio	_{0.277 GB}	_{English, Italian, Spanish}	_{Categorical vs Dimensional Perception of Italian Emotional Speech}	_Restricted	_{EULA: End User License Agreement}
_ANAD	₂₀₁₈	_{1384 recording by multiple speakers.}	_{3 emotions: angry, happy, surprised.}	_Audio	_{2 GB}	_Arabic	_{Arabic Natural Audio Dataset}	_Open	_{CC BY-NC-SA 4.0}
_EmoSynth	₂₀₁₈	_{144 audio file labelled by 40 listeners.}	_{Emotion (no speech) defined in regard of valence and arousal.}	_Audio	_{0.1034 GB}	_–	_{The Perceived Emotion of Isolated Synthetic Audio: The EmoSynth Dataset and Results}	_Open	_{CC BY 4.0}
_CMU-MOSEI	₂₀₁₈	_{65 hours of annotated video from more than 1000 speakers and 250 topics.}	_{6 Emotion (happiness, sadness, anger,fear, disgust, surprise) + Likert scale.}	_{Audio, Video}	_{190.1 GB}	_English	_{Multi-attention Recurrent Network for Human Communication Comprehension}	_Open	_{CMU-MOSEI License}
_VERBO	₂₀₁₈	_{14 different phrases by 12 speakers (6 female + 6 male) for a total of 1167 recordings.}	_{7 emotions: Happiness, Disgust, Fear, Neutral, Anger, Surprise, Sadness}	_Audio	_–	_Portuguese	_{VERBO: Voice Emotion Recognition dataBase in Portuguese Language}	_Restricted	_{Available for research purposes only}
_CMU-MOSI	₂₀₁₇	_{2199 opinion utterances with annotated sentiment.}	_{Sentiment annotated between very negative to very positive in seven Likert steps.}	_{Audio, Video}	_{4.3 GB}	_English	_{Multi-attention Recurrent Network for Human Communication Comprehension}	_Open	_{CMU-MOSI License}
_MSP-IMPROV	₂₀₁₇	_{20 sentences by 12 actors.}	_{4 emotions: angry, sad, happy, neutral, other, without agreement}	_{Audio, Video}	_{3.4 GB}	_English	_{MSP-IMPROV: An Acted Corpus of Dyadic Interactions to Study Emotion Perception}	_Restricted	_{Academic License & Commercial License}
_CREMA-D	₂₀₁₇	_{7442 clip of 12 sentences spoken by 91 actors (48 males and 43 females).}	_{6 emotions: angry, disgusted, fearful, happy, neutral, and sad}	_{Audio, Video}	_{0.607 GB}	_English	_{CREMA-D: Crowd-sourced Emotional Multimodal Actors Dataset}	_Open	_{Open Database License & Database Content License}
_{Example emotion videos used in investigation of emotion perception in schizophrenia}	₂₀₁₇	_{6 videos:Two example videos from each emotion category (angry, happy and neutral) by one female speaker.}	_{3 emotions: angry, happy and neutral.}	_{Audio, Video}	_{0.063 GB}	_English	_–	_Open	_{Permitted Non-commercial Re-use with Acknowledgment}
_EMOVO	₂₀₁₄	_{6 actors who played 14 sentences.}	_{6 emotions: disgust, fear, anger, joy, surprise, sadness.}	_Audio	_{0.355 GB}	_Italian	_{EMOVO Corpus: an Italian Emotional Speech Database}	_Open	_–
_RECOLA	₂₀₁₃	_{3.8 hours of recordings by 46 participants.}	_{negative and positive sentiment (valence and arousal).}	_{Audio, Video}	_–	_–	_{Introducing the RECOLA Multimodal Corpus of Remote Collaborative and Affective Interactions}	_Restricted	_{Academic License & Commercial License}
_{GEMEP corpus}	₂₀₁₂	_{Videos10 actors portraying 10 states.}	_{12 emotions: amusement, anxiety, cold anger (irritation), despair, hot anger (rage), fear (panic), interest, joy (elation), pleasure(sensory), pride, relief, and sadness. Plus, 5 additional emotions: admiration, contempt, disgust, surprise, and tenderness.}	_{Audio, Video}	_–	_French	_{Introducing the Geneva Multimodal Expression Corpus for Experimental Research on Emotion Perception}	_Restricted	_–
_OGVC	₂₀₁₂

README 内容过长，已截取前 64 KB。完整内容请前往 GitHub 仓库查看。

暂无评论

SER-datasets