UCF-Crime-DVS
收藏资源简介:
UCF-Crime-DVS是由宁波大学 Faculty of Electrical Engineering and Computer Science创建的第一个面向视频异常检测的大型事件相机数据集。该数据集包含1900个事件流,涵盖13种异常类别,与UCFCrime数据集保持一致。每个视频平均时长为242秒,总时长共计128小时,具有高分辨率1280×720。数据集通过事件相机捕捉,能够提供丰富的动态信息和高时间分辨率,适用于视频异常检测任务。
UCF-Crime-DVS is the first large-scale event camera dataset for video anomaly detection, developed by the Faculty of Electrical Engineering and Computer Science of Ningbo University. This dataset contains 1900 event streams, covering 13 anomaly categories consistent with the UCFCrime dataset. Each event stream has an average duration of 242 seconds, with a total duration of 128 hours and a high resolution of 1280×720. Captured via event cameras, this dataset offers rich dynamic information and high temporal resolution, making it applicable to video anomaly detection tasks.

- 1UCF-Crime-DVS: A Novel Event-Based Dataset for Video Anomaly Detection with Spiking Neural Networks宁波大学 Faculty of Electrical Engineering and Computer Science · 2025年



