Research of a Real-time Hand Tracking Algorithm

Gesture analysis is an important region in HCI (human computer interaction) technology and hand tracking is a key step of gesture analysis. A modified CAMSHIFT (continuously adaptive mean shift) algorithm, which can track hand in real time, is presented in this paper. Considering the hand motion and skin color probability distribution all together, the modified algorithm, which enhances anti-noise ability, avoids distracters and deals well with shelter problem, is put forward. Tracking precision is improved. Experiment results show this algorithm is fast, accurate, and robust. It can be used in other region tracking special color