Limb Activity and Recreational Action Posture Video Dataset

In the modern medical and health field, with the enhancement of public health awareness and the arrival of an aging society, physical activity monitoring and rehabilitation training have become crucial. However, existing solutions often lack the ability to accurately identify and analyze complex limb movements, especially in home environments. The Limb Activity and Recreational Action Posture Video Dataset aims to improve the accuracy of such systems to better support users' self-health management and remote monitoring by healthcare professionals. The video collection in this dataset uses high-definition camera equipment to record different limb activities in various living environments. Quality control is reflected in multiple rounds of annotation and consistency checks, with the annotation team consisting of 20 experts in the medical field and motion capture. The data is pre-processed, including editing, denoising, and frame alignment, and is finally organized and stored in MP4 format to facilitate rapid distribution and use.