A Human Target Detection and Tracking Method Based on Adaptive Difference and GVF-Snake Algorithm

摘要

Detection of Frame difference fails when the human target is stationary in course of moving, this paper presents a method based on combination of adaptive difference and GVF-snake algorithm to solve it. Adaptive differential detection algorithm can accurately extract the target contour, and use it as the initial contour of GVF-snake model which cannot automatically extract it after we got the target. In the process of detection and tracking, calculating GVF field of the whole picture consume too much time, so we use the method of sub-region to improve the real-time. The experimental results show that, the algorithm can provide the actual body contour for GVF-snake model, and effectively track whether the target is stationary or moving.