An OPS Hand Tracking Algorithm Based on Optical Flow

In this paper, a real-time technique for segmenting a hand, determining hand´s locations is described. The algorithm consists of two steps; the first is to location the hand area in noise condition and segment the part of hand. Then seek for angles of the fingers and location of fingertip with a novel factor determination in special OPS (optical parameter sequence). It keeps the advantages of optical flow and overcome the complex computation cost and recognizes the fingerspell correctly. According to the traditional technique, the algorithm has a good performance.