Abstract: Face recognition is emerging as one of the popular field in the biometric research. It is used in various surveillance systems for the security purpose as it doesn’t need object comparison. The main advantage of using face recognition system is its uniqueness and acceptance over other biometric systems. Though this system is considered to be accurate but the detection of the face is a difficult process due to the high degree of variability in faces. Face recognition is a section of pattern recognition in which human visual perception is saved in the computer. Many researchers are working on this field for many of the years, many algorithms and techniques are developed to update the traditional systems. Some of the techniques that are common these days are PCA, LDA, and Gabor etc. But these approaches individually are not that much efficient in some of the cases. In this research work a new approach is proposed, in which the main concentration is to extract the features with the PCA and LDA techniques as well as the combined systems. This is a much better approach to work with, as suggested from the literature and also more successful for the large dataset. After combining both of the feature extraction, finally the classification will be done and the performance will be evaluated.
Keywords: Face Recognition, Biometric System, Linear Discriminant Analysis, Principal Component Analysis.