Multiple Faces Detection Through Facial Features and Modified Bayesian Classifier

A new multiple faces detection method based on facial features is proposed, which gets the face candidates with the help of skin color and makes use of wavelet express of images and the Principal Component Analysis (PCA) to obtain the eigenvectors distinguishing faces and non-faces, and modifies Bayesian classifier to detect multiple faces of input images. ω, the parameter of the modified rules, could control detection accuracy and error rate to be applied to different application by setting its different values. In addition, after classification, a mosaic template is used to exclude fake faces to ensure high accuracy and low error probability.