Authors : Ramesh Naidu, Jagadeesh H S and Ramachandra A C
The face recognition system attains good accuracy with a large set of training images. However, with a smaller training set their performance is relatively poor. In this paper, a face recognition method is explored that achieves good results when only a small training set is available. The proposed model is based on Image Average with Histogram Equalization (IAHE) and Global Binary Pattern (GBP). The face image is preprocessed using a proposed scheme IAHE to highlight the facial features and then GBP is applied to extract unique facial features. The features of test face image are compared with that of database images using Euclidian distance. It is observed that the results are improved considerably on different public databases by varying number of images used for training.
Keywords: Biometrics, Histogram Equalization, Face Recognition