VGG-Sound
收藏arXiv2025-09-30 收录
下载链接：
http://www.robots.ox.ac.uk/~vgg/data/vggsound/
下载链接
链接失效反馈官方服务：
资源简介：
该数据集名为VGGSound，是一个大规模的视频数据集，包含了309个类别，涵盖了日常生活中广泛的音频事件。每个视频的时长为10秒，旨在用于进行音频事件分类的任务。
This dataset, named VGGSound, is a large-scale video dataset that comprises 309 categories covering a wide range of daily audio events. Each video has a duration of 10 seconds, and it is designed for audio event classification tasks.
