Face detection from cluttered images using Gabor filter features

This paper proposes a classification-based approach using Gabor filter features for detecting faces in clutter images. The underlying classifier is a polynomial neural network (PNN) which is a single layer network performing nonlinear classification by using the polynomial expansion of pattern features as the network input. The features based on Gabor filters extracted from local image are applied to be the input of the classifier. The dimensionality of the Gabor feature vector is reduced by the principal component analysis (PCA). The feasibility of the proposed method has been proven by experimental results on testing a large number of images and the comparison to several state-of-the-art approaches.