Optimized fuzzy system using genetic algorithm to detect faces in color images

A human face detection method for color images is presented in this paper. The system is composed of three subsystems: skin color segmentation, lip color segmentation and face blobs selection subsystem. Whole these algorithms are fuzzy rule base ones, which are designed empirically, and then optimized by genetic algorithm. In the first stage, skin color regions are selected in the input image. Within each of the skin area, lip pixels are searched using second subsystem, and applied as a feature to identify face candidates in the skin regions. Utilizing the lip area and position relative to the skin area, and face shape information, the third subsystem is materialized to choose face blobs. To precise evaluation of the proposed system, the false positive and false negative of each subsystem, are reported for the empirically designed system as well as the optimized system. Obtained results show a remarkable decrease in false positive and false negative for optimized algorithms compared to empirically designed ones. Finally, 98% detection rate is achieved using proposed method.