BanSpEmo

BanSpEmo是一个专为孟加拉语设计的情感语音识别数据集，由常州大学和孟加拉国格林大学合作创建。该数据集包含792个音频记录，总时长超过1小时23分钟，涵盖了六种基本情感：厌恶、快乐、悲伤、惊讶、愤怒和恐惧。数据集由22名母语为孟加拉语的演讲者录制，每种情感类别中音频记录数量均衡。BanSpEmo旨在推动孟加拉语情感和语音识别研究及相关应用的发展，适用于心理学检查、机器人技术、移动服务等多个领域。BanSpEmo is a sentiment speech recognition dataset specifically designed for the Bengali language, jointly created by Changzhou University and Green University of Bangladesh. This dataset contains 792 audio recordings with a total duration of over 1 hour and 23 minutes, covering six basic emotions: disgust, happiness, sadness, surprise, anger, and fear. Recorded by 22 native Bengali speakers, the dataset has a balanced number of audio samples across each emotion category. BanSpEmo aims to promote the development of Bengali sentiment and speech recognition research and related applications, and is applicable to multiple fields such as psychological assessment, robotics, mobile services and other domains.