Video-Based Fall Detection Dataset with 2017 Activities from 29 Subjects

This dataset comprises a diverse collection of 2,017 video recordings for fall detection, captured from various subjects. It includes 999 falls and 1,017 Activities of Daily Living (ADL), providing a balanced and realistic representation of real-world scenarios. Fall events cover a range of conditions, including falls to the left, right, front, and back, as well as complex situations such as falling while attempting to sit or from elevated positions. ADL activities include walking, running, standing up from the ground, and other daily movements. This dataset serves as a valuable resource for developing and evaluating advanced video-based fall detection models.本数据集包含面向跌倒检测的多样化录制视频共2017段，采集自不同受试者。该数据集涵盖999段跌倒事件视频与1017段日常活动（Activities of Daily Living, ADL）视频，能够真实且均衡地还原现实场景中的各类情况。跌倒事件涵盖多种类型，包括向左、向右、向前、向后跌倒，以及诸如试图坐下时跌倒或从高处跌落等复杂场景。日常活动则包含行走、奔跑、从地面起身及其他日常肢体动作。本数据集可作为研发与评估先进视频式跌倒检测模型的宝贵资源。