360-VSumm

360-VSumm数据集是由信息技术和工业研究所创建，旨在支持360度视频摘要技术的研究和评估。该数据集包含多种主题的视频内容，如音乐表演、体育赛事、短片和纪录片，涵盖室内外环境，使用静态或移动摄像机拍摄。数据集中的视频包含多个时间重叠或并行的事件，为开发通用360度视频摘要方法提供了丰富的资源。通过人工标注，该数据集提供了视频摘要的基准，用于训练和客观评估视频摘要方法，特别适用于解决传统设备上消费360度视频内容的问题。

The 360-VSumm dataset was created by the Institute of Information Technology and Industry to support research and evaluation of 360-degree video summarization techniques. This dataset contains video content across various topics, including musical performances, sporting events, short films, and documentaries. It covers both indoor and outdoor environments, and is shot with either static or mobile cameras. The videos in the dataset include multiple temporally overlapping or parallel events, providing a rich resource for developing general-purpose 360-degree video summarization methods. Through manual annotation, this dataset provides a benchmark for video summarization, which is used for training and objectively evaluating video summarization methods, and is particularly suitable for addressing the challenges of consuming 360-degree video content on traditional devices.