Efficient face and gesture recognition techniques for robot control

This paper presents a visual recognition system for interactively controlling a mobile robot. First, the robot identifies the operator by human facial recognition, and then determines the actions analyzing the human hand gestures. For facial recognition, the adaptive region-growing algorithm is proposed to estimate the location of face region. A genetic algorithm is then applied to search for the accurate facial feature positions. For gesture recognition, we use adaptive color segmentation, hand finding and labeling with blocking, morphological filtering, and gesture actions are found by template matching, and skeletonizing. The results show 95% correct recognition ratio compared to less than 90% mentioned in other papers.