Face Recognition using Nearest Neighbour and Nearest Mean Classification Framework : Empirical Analysis, Conclusions and Future Directions

Abstract

Abstract Human Face recognition algorithms have made huge progress in the last decade. In this manuscript, we have presented an approach for the implementation of a face recognition system in a successful manner by varying pose, scale, lighting, and age variation. The different empirical analysis was performed with various datasets for face detection and face identification. Face identification system detects efficiently segments and recognizes face in a cluttered sequence under varying pose, lighting and age variations. From this experimental analysis morphological model outperformed k-NNC, NMC based closest mean classifier and informative knowledge distillation with fairly reasonable accuracy. Three proposed methods on the basis of an efficient way of handling the face recognition problems. The morphological method outperformed well when compared with k-NNC, NMC based closest mean classifier a proposed method, and another innovative method named Informative knowledge Distillation. The morphological method is suitable for large datasets where occlusion, pose variation, age variations, and different expression of images.