AGS

AGS数据集是由哈尔滨理工大学计算机科学与技术系开发的，专注于家庭环境声音的数据集。该数据集包含3986条声音记录，来源于972个视频文件，涵盖65个类别，旨在解决现有声音数据集在家庭环境声音场景中的不足。创建过程中，数据集通过视频和音频的综合分析，提取关键声音片段并进行标注。AGS数据集的应用领域包括环境声音场景理解和声音事件识别，特别是在智能监控和活动识别等场景中具有重要价值。

The AGS dataset, developed by the Department of Computer Science and Technology, Harbin University of Science and Technology, is a collection focused on household environmental sounds. It contains 3986 audio recordings derived from 972 video files, covering 65 categories, and aims to address the shortcomings of existing audio datasets in household environmental sound scenarios. During its development, key audio segments were extracted and annotated through comprehensive analysis of both video and audio materials. The AGS dataset has applications in environmental sound scene understanding and sound event recognition, and holds significant value especially in scenarios such as intelligent monitoring and activity recognition.