Rights Creative Commons: Attribution 3.0 Hong Kong License REAL TIME HUMAN BODY TRACKING USING WAVENET

Human body tracking is useful in applications like med-ical diagnostic, human computer interface, visual surveil-lance etc. In this paper, a trajectory-learning algorithm us-ing wavenet is proposed to track human body in real time without sacrificing accuracy. Human body is located within a small searching window using color and shape as heuris-tic. The location and size of the searching window are es-timated using the proposed wavenet-based algorithm. The effectiveness and efficiency of the proposed tracking system were tested. The results show that the proposed system can track the human body in real time under various conditions including clutter background, background with distractor, and scene with object occlusion. 1