Multi-feature Fusion Method of Driver Face Location Based on Area Coincidence Degree and Prior Knowledge

Speedy and reliable face location is a key for monitoring driver fatigue driving condition using machine vision methods. According to the disadvantage of face location methods based on single feature in accuracy and reliability at present, first, an improved face location method based on Haar-like feature and Adaboost is used to detect the possibly existing face region in the whole image, then the detected region is extended properly and a face location method based on skin color feature in normalized rgb and HSV color spaces is used to locate the face again in the extended area, finally, fusion location of driver face region is achieved by the defined area coincidence degree and prior knowledge of human face. Experiments carried out in various road environments demonstrate the validity of the fusion method proposed.