Wavelet Based Face Recognition Using Statistical Modeling

Face recognition is one of the challenging applications of image processing. Robust face recognition algorithm should posses the ability to recognize identity despite many variations in pose, lighting and appearance. Nowadays, Principal Component Analysis (PCA) has been widely adopted as the potential face recognition algorithm. However, it has limitations like poor discriminatory power and large computational load. In view of these limitations, this paper proposes a new approach in using PCA on wavelet sub bands. In the proposed method, wavelet transform is used to decompose an image into different frequency sub bands, and a mid-range frequency sub band is used for PCA representation. In comparison with the conventional use of PCA, the proposed method gives better recognition rate and discriminatory power. Further, the proposed method reduces the computational load significantly even when the image database is large. This paper details the design and implementation of the proposed method, and presents encouraging experimental results with standard Face Database.