Performance Evaluation of Face Recognition Algorithms

Biometric-based techniques have emerged for recognizing individuals instead of using passwords, PINs, smart cards, plastic cards, tokens etc for authenticating people. Automated face recognition has become a major field of interest. In this field several facial recognition algorithms have been explored in the past few decades. A face recognition system is expected to identify faces present in images and videos automatically. The input to the facial recognition system is a two dimensional image, while the system distinguishes the input image as a user’s face from a pre-determined library of faces. Finally, the output is a discerned face image. This paper deals with the comparison of two popular dimensionality reduction algorithms such as PCA and LDA. Here, our main goal is to evaluate the performance of Principal Component Analysis and Linear Discriminant Analysis for large training data set. Finally, we concluded that LDA outperforms PCA for the large samples of training set. KeywordsPrincipal Component Analysis, Linear Discriminant Analysis, Eigenfaces, ORL, Face recognition. _________________________________________________*****_________________________________________________