Detection of Human Motion Using a Fusion Method Based on Optical Flow and Skin Color

Optical flow algorithm has its difficultly in detecting the whole human region in the complex environment. An integrated method of detecting human motion based on optical flow and skin color is proposed. The motion region is detected by optical flow computation. Skin detection then concurred by computing the Mahalanobis distance. A whole object is detected by a fusion method based on optical flow and skin color. Experimental results showed that the method permits the detection of the whole human object.