FACE RECOGNITION: A NEW FEATURE SELECTION AND CLASSIFICATION TECHNIQUE

Abstract

A novel feature selection and classification technique for face recognition is presented this paper. Genetic Algorithms (GAs) for feature selection and Artificial Neural Network (ANN) for classification are incorporated in the proposed technique. The proposed GAs-ANN technique has two purposes in this research: 1) fusion and selection of features for face recognition. 2) location of significant areas inside ach facial region. Facial regions are identified by using distance threshold method based on center coordinate information of each facial region. The average grey level value feature is extracted from each facial region. Then these features are combined to form the input feature vector for the GAs-ANN technique. A set of experiments is conducted on a subset of the FERET database. The results show that the proposed approach is promising. A comnrehensive comparison with other existing face recognition approaches is included.