Sitting Posture Recognition in Real-Time Combined with Index Map and BLS

Wanting Sun, Ze Zhou, Hongjun Li

Abstract: The development of modern science and technology has seriously changed people's living and working habits. Bad sitting habit undoubtedly has an important impact on human health. Therefore, this paper proposes a real-time sitting posture recognition algorithm based on index map and BLS model. Firstly, use Kinect to collect body index maps and build self-built sitting posture database; then use SSIM to supervise the change of sitting posture, the BLS is used to recognize the category of sitting posture in real-time; finally, as for the excessive frames in video, this paper proposes double thresholds cascade algorithm to detect excessive frames. The experimental results show that the algorithm proposed can effectively recognize 8 types of sitting postures with the accuracy of 98.58%, the speed of 162fps and effectively detect the excessive frames.