Dynamic Gesture Capture, Location and Tracking Based on MeanShift Algorithm

摘要

In this paper, dynamic gesture capture and tracking based on MeanShift algorithm for human-computer interaction system is proposed. In this method, the rectangular box representing the hand position is used as the control signal input and the initial detection target tracking. Thus the dynamic gesture video is effectively processed. In this paper, Python3 OpenCV library is adopted as the main processing library for experimental analysis. The initial capture and positioning of hand position and subsequent tracking experiments are carried out on a video of hand movement. Through the comparison and analysis with Camshift algorithm, the experimental results show that MeanShift algorithm achieves a good dynamic tracking effect of gesture position.