A Novel Approach to Face Recognition using Split PCA

Deepali H. Shah, Dr. J. S. Shah

Abstract Human face recognition has attracted significant consideration as one of the most effective applications of image analysis and understanding. Face recognition provides us a convenient way to identify and recognize a person in a large database. Lighting condition changes, facial expressions, and pose variations complicate the face recognition. This paper will present a novel approach for solving face recognition problem. Our method modifies Principal Component Analysis (PCA) method as Split PCA and uses Support Vector Machine (SVM) classifier with constant hyperplane. The experiments of the proposed method have been conducted on ORL data sets. The experiment result is presented and discussed, which shows the effectiveness of the strategy described. Keywords: PCA, SVM, hyperplane, eigenfaces

Title:	A Novel Approach to Face Recognition using Split PCA
Author Name:	Deepali H. Shah, Dr. J. S. Shah
Abstract:	Abstract Human face recognition has attracted significant consideration as one of the most effective applications of image analysis and understanding. Face recognition provides us a convenient way to identify and recognize a person in a large database. Lighting condition changes, facial expressions, and pose variations complicate the face recognition. This paper will present a novel approach for solving face recognition problem. Our method modifies Principal Component Analysis (PCA) method as Split PCA and uses Support Vector Machine (SVM) classifier with constant hyperplane. The experiments of the proposed method have been conducted on ORL data sets. The experiment result is presented and discussed, which shows the effectiveness of the strategy described. Keywords: PCA, SVM, hyperplane, eigenfaces
Cite this article:	Deepali H. Shah, Dr. J. S. Shah , " A Novel Approach to Face Recognition using Split PCA " , International Journal of Emerging Trends & Technology in Computer Science (IJETTCS), Volume 4, Issue 3, May - June 2015 , pp. 130-132 , ISSN 2278-6856.
Full Text [PDF] Home