A Face Detection Based On Fuzzy Set Theoretics

In the recent years, the subject of face detection which was originally put forward as location part has been placed a lot of importance as an independent subject by researchers due to the practical value in the fields of safe access control, vision inspection, content-based retrieval and new generation human-machine interface.Human-face detection is a technique which can find the positions and the sizes of human faces in color images through some algorithms. It is the first step in human face detection, human face tracking, posture estimation and expression identification. And the study of human sexual identification based on human face characters is a hotspot. Its search fruit will play an important part in sexual analysis of humans in some public areas. There are some shortages in some old methods of human face detection at the sides of speed or nicety rate, and sexual identification is a new studying hotspot. So it will be important to excogitate a fast and exact human face detection method and an effective method of sexual identification.This paper analyses the feasibility that fuzzy set theoretics apply to face detection. It trains the sample set by Haar-like feature and membership function, and selects appropriate weak classifiers through the entropy of feature set and AdaBoost learning algorithm. Then distribution face detector is established. In the detection process, detector can rapidly wash out the sub-window which is unlike face through the front of simple stronger classifiers; for the sub-window which is like face, in terms of similar degree distributer can dynamically select the back of stronger classifiers to determine whether it is a face. This face detector has been tested on the MIT+CMU frontal face test set, the results show that it can effectually improve detection efficiency in the condition of detection performance reducing not too much.