HARDVS

该数据集是一个大规模的事件相机用于人类活动识别的基准数据集，涵盖了300个类别，并包含了超过10万条事件序列。该数据集包含了多种挑战性场景，如多视角、多光照、多运动、动态背景以及遮挡情况。数据集是通过DAVIS346相机收集的，旨在充分展现DVS传感器在复杂环境下的独特特性。该数据集规模宏大，拥有超过107,646个视频序列，其任务是进行人类活动识别（Har）。

This is a large-scale benchmark dataset for human activity recognition (HAR) using event cameras. It covers 300 categories and contains over 100,000 event sequences. The dataset includes various challenging scenarios such as multi-view perspectives, varying illumination conditions, diverse motion patterns, dynamic backgrounds, and occlusion cases. Collected with the DAVIS346 camera, this dataset is designed to fully showcase the unique characteristics of Dynamic Vision Sensor (DVS) in complex environments. With an extensive scale of more than 107,646 video sequences, its core task is human activity recognition (HAR).