Discrete region inferred differential evolution feature selection architecture for human face recognition

Abstract

Face Recognition has received a lot of attention in wide variety of the security application employing face based biometric authentication due to its non intrusive data acquisition. Especially artificial intelligence play important role in face recognition. Numerous researches have been carried out using machine learning technique in order to accuracy of the recognition. In this paper, a novel technique titled as Discrete Differential Evolution (DDE) architecture has been proposed. The proposed architecture will employed after set of the primary image processing steps such as preprocessing, feature extraction and feature selection. Initially preprocessing of the images is performed with image normalization technique and image enhancement technique. Preprocessed image will explored to Linear Discriminant Analysis technique to extract the features of the image. Extracted features will contain some redundant and irrelevant information of the images, it has to eliminate and optimal feature has to be selected to the recognition task. Discrete Region Inferred Differential Evolution architecture is employed to feature selection. Extracted features will be consider as search space, objective function of the DDE generates the optimal features solutions with respect to fitness criteria. Finally recognition task is performed with optimal features using support vector machine.