Human Face Identification using Moments and Transformations

Abstract

Abstract Face recognition one of the most promising techniques such that the importance of biometric user identification is increasing every day, several methods have been suggested to perform it such as classification, deep learning, statistical, moments, etc., this work describes different approaches to develop biometric technique, based on the moments and transformation. The proposed method for face recognition is based on Legendre moments and singular value decomposition for feature extraction. Local approaches by dividing images into several blocks (overlapped blocks and non-overlapped) have been adopted to gain the best recognition rate, as well as normalization using the Z-score method, which has been used. The outcomes experimental showed that the suggested system is effective, it has been tested using ORL face image databases with 10 cases and achieved a recognition rate from 85.27-100%, also, applied on FEI Brazil face database with 10 cases and achieved a recognition rate from to 79.77-100%.