Fast Face Detection Based on Fuzzy Set Theory

In order to improve efficiency of face detector, fuzzy set theory is used in establishing distribution face detector. This detector trains the sample set by Haar-like feature and membership function, and selects appropriate weak classifiers through the feature setpsilas entropy and AdaBoost learning algorithm. Subsequently distribution face detector is established and tested on the MIT+CMU frontal face test set. The results show that the detector can rapidly eliminate the sub-window which is unlike a face through the front simple stronger classifiers; and that distributor can dynamically select the back stronger classifiers to determine whether it is true or not in terms of similar degree when the image sub-window looks like a face. This detector can effectually improve detection efficiency in the condition of detection performance reducing not too much.