A Real Time Face Detection Method in Human-Machine Interaction

To solve the problem of detecting single face in human-machine interaction, We propose a simple yet new real time face detection method which is based on the individual chrominance discrepancy of skin and lip. First, find out skin area based on a threshold value in HSV color model, then search the rough lip area based on the skin chrominance adaptively. Then processed with R/G filter Lip was found, based on the geometrical relation between face and lip Face was finally located. Experiments on service robot demonstrate that the method is effective, fast compared with other methods and fit for detecting and tracking face in real time application.