Towards an efficient face recognition approach

As modern technology evolves, the use of face recognition system is scattering in different sectors of commercial markets rather than in security purposes only. Various approaches are introduced for face recognition system, among them principal component analysis is one of the simplest and efficient method. To improve the performance of face recognition, choosing a threshold value and minimum number of principal components are important. This paper proposes a new face recognition method which uses smaller number of principal components to retain the maximum variance for an optimized solution. With this optimization, the proposed method shows better accuracy for facial recognition than existing dominant work among a huge set of face images in real time face images.