Scholarlink

<p>We present the first sizable corpus of Thai speech emotion recognition, THAI-SER, comprising 41&#xa0;h and 36&#xa0;min (27,854 utterances) across 100 recordings recorded in different environments: Zoom and two studio setups. The recordings contain both scripted and improvised sessions, acted by 200 professional actors (112 females and 88 males, aged 18 to 55) and were directed by professional directors. There are five primary emotions: neutral, angry, happy, sad, and frustrated, assigned to the actors when recording utterances. The utterances are annotated with an emotional category using crowdsourcing. To quality control the annotation process, we also design an extensive filtering and quality control scheme to ensure that the majority agreement score remains above 0.71. We evaluate our annotated corpus using two metrics: inter-annotator reliability and human recognition accuracy. Inter-annotator reliability score was calculated using Krippendorff’s alpha, where our corpus, after filtering, achieved an alpha score of 0.692, higher than the recommended score of 0.667. For human recognition accuracy, our corpus scored up to 0.772 post-filtering. We also provide the results of the model trained on the corpus evaluated on both in-corpus and cross-corpus setups. The corpus is publicly available under a Creative Commons BY-SA 4.0 (On Github Release - <a href="https://github.com/vistec-AI/dataset-releases/releases/tag/v1">https://github.com/vistec-AI/dataset-releases/releases/tag/v1</a> On HuggingFace - <a href="https://huggingface.co/datasets/airesearch/thai-ser">https://huggingface.co/datasets/airesearch/thai-ser</a>) as well as our codes for the experiments (<a href="https://github.com/tann9949/thaiser-experiments">https://github.com/tann9949/thaiser-experiments</a>).</p>

错误：搜索内容不能为空，请输入英文关键词
错误：关键词超出字数限制，请精简
高级检索

Thai speech emotion recognition (THAI-SER) corpus

  • 期刊论文

  • 发表日期：

    2026年7月24日

查看全文
  • Jilamika Wongpithayadisai,
  • Chompakorn Chaksangchaichot,
  • Soravitt Sangnark,
  • Patawee Prakrankamanant,
  • Krit Gangwanpongpun,
  • Siwa Boonpunmongkol,
  • Premmarin Milindasuta,
  • Dangkamon Na-Pombejra,
  • Sarana Nutanong,
  • Ekapol Chuangsuwanich

摘要

We present the first sizable corpus of Thai speech emotion recognition, THAI-SER, comprising 41 h and 36 min (27,854 utterances) across 100 recordings recorded in different environments: Zoom and two studio setups. The recordings contain both scripted and improvised sessions, acted by 200 professional actors (112 females and 88 males, aged 18 to 55) and were directed by professional directors. There are five primary emotions: neutral, angry, happy, sad, and frustrated, assigned to the actors when recording utterances. The utterances are annotated with an emotional category using crowdsourcing. To quality control the annotation process, we also design an extensive filtering and quality control scheme to ensure that the majority agreement score remains above 0.71. We evaluate our annotated corpus using two metrics: inter-annotator reliability and human recognition accuracy. Inter-annotator reliability score was calculated using Krippendorff’s alpha, where our corpus, after filtering, achieved an alpha score of 0.692, higher than the recommended score of 0.667. For human recognition accuracy, our corpus scored up to 0.772 post-filtering. We also provide the results of the model trained on the corpus evaluated on both in-corpus and cross-corpus setups. The corpus is publicly available under a Creative Commons BY-SA 4.0 (On Github Release - https://github.com/vistec-AI/dataset-releases/releases/tag/v1 On HuggingFace - https://huggingface.co/datasets/airesearch/thai-ser) as well as our codes for the experiments (https://github.com/tann9949/thaiser-experiments).