Face Detection Based On Template-matching, Mosaic Image And Support Vector Machines

Face detection is the process of recognizing a pattern in an image as a face pattern and subsequently segmenting it from the rest of the image. Human face detection is a challenging exercise because the face can have varying pose, size, skin color, and facial expression. In addition, other factors, such as wearing of glasses, presence or absence of hair, and occlusion can make the appearance of the face unpredictable. Other effects, such as varying lighting conditions and complexity of the scene containing the face can make the generalization of face detection algorithms difficult.In this paper ,we provide a new human face detection algorithm based on template-matching, mosaic image and Support Vector Machines. This algorithm is mainly composed of three sections, which are image preprocessing ,production and selection of eye-dots, selection of face. Image preprocessing includes transformation of image format, scaling of image size , smooth and noise -removing , gray level distribution normalization. In the process of production and selection of eye-dots, Firstly, the binary image contained human eye region is abstracted by morphological translation , labeled based on connection and the labeled region are assembled into candidates of human eye dots . After the dots are fine tuning based on gray projection calculation , they are chosen ground on gray and edge information. The combinations of two of the dots form the candidates of human faces. Three sieve methods are used in the selection of human faces. The first method is based on general knowledge. The second one is based on template-matching and mosaic image. Last, the candidate of faces will be passed to the SVM (support vector machines) classifier for the final decision.Experimental results show that our algorithm has some detection effect on face image on simple background.