Local Binary Pattern and PCA Approaches: Towards for Developing Face recognition system

Abstract

`Face recognition is one of biometrics system used for surveillance   purpose.  It use to   discovery criminals, suspected terrorists   and missed children. The face recognition is term to identify human by using algorithms. In this paper,   Local Binary Pattern (LBP)   approach   has applied to extract features for face recognition. LBP approach is one of powerful and robust method for extraction features from face. Main contributions of this present research work   are:   First   LBP    approach   used to extract   importance features from face.  Second:  Feature selection methods, Principal Component Analysis (PCA)   has employed to remove irrelevant features    for increase   accuracy of classification. For face recognition the   most significant features       are necessary    due to   the face has more dimensionality. Third,   Classification,   Two classifiers are applied for recognition purpose.  Support Vector Machine (SVM) and Linear Discriminates (LD) algorithms    have   used   to classify the features   vector which has obtained by LBP method for face recognition. For  experimental  analysis ,  the   Olivetti Research Laboratory  (ORL ) standard data    has  been  used  to evaluate  the proposed system. The empirical   analysis results   of proposed system show   that it is better in terms   of accuracy performance measures. A Comparative   analysis results between the classifiers with all features and with PCA method is presented.  It is observed that    the performance of classifiers with whole features is   better. However, the classifier with   PCA   method is better in cost of time. Finally, it concluded that the   proposed   system   is   more robust and better    for face recognition.