Face Recognition Using Computational Efficient Algorithms

Research in pattern recognition and biometrics have received considerable attention in recent years among computer vision communities, but have remained very problematic in actual applications. This paper performed a detail and direct comparison of two algorithm, i.e. Principal Component Analysis (PCA) and Fisher Linear Discriminant (FLD), for face recognition. PCA project the image onto the first few principal component, where foremost principal component is a direction that maximizes the variance of projected image and last principal component captures least amount of variance among images. Even when the light and facial expressions are greatly varied, FLD creates a group of images of the same class and separately of different class. The eigen vectors of face datasets are used as a significant features for classification. Here six facial databases (FEI face database, Face94’ Essex face database, CASIA, FERRET database and NITH face database) have been taken for comparison of performance. These six databases provide a comprehensive dataset for testing the performance of the PCA and FLD algorithm.