VOTE400

VOTE400是由电子通信研究所和MINDs Lab公司合作创建的大型韩语语音数据集，专注于老年人的语音分析和识别。该数据集包含300小时的连续对话语音和100小时的朗读语音，均由65岁以上的老年人录制。数据收集过程中，得到了韩国政府机构Ji-Won Center的支持，通过其支持的Saa-Raang-It-Gi项目收集了自发对话语音，并通过专门的语音收集系统收集了朗读语音。VOTE400旨在提升老年护理机器人的语音识别性能，解决现有商业语音识别系统在老年人语音识别上的不足。

VOTE400 is a large-scale Korean speech dataset co-created by the Electronics and Communications Research Institute and MINDs Lab, Inc., focusing on speech analysis and recognition for the elderly. The dataset contains 300 hours of continuous conversational speech and 100 hours of read speech, all recorded by seniors aged 65 and above. During the data collection process, it received support from the South Korean government agency Ji-Won Center. Spontaneous conversational speech was collected through its supported Saa-Raang-It-Gi project, while read speech was collected via a dedicated speech collection system. VOTE400 aims to improve the speech recognition performance of elderly care robots, addressing the shortcomings of existing commercial speech recognition systems in recognizing elderly speech.