Principal Component Analysis Based on Discrete Wavelet Transform

Abstract

The contribution of this paper is to improve the face recognition rate by applying two-dimensional Discrete Wavelet Transform (2DWT) subband coefficients to reduce the high dimensional image into low dimensional image. The Principal Component Analysis (PCA) is being applied to find the face recognition accuracy rate, using ORL image database. A comparison between the PCA, first subband level 2DWT/PCA and second subband level 2DWT/PCA have being evaluated according to the recognition accuracy, dimensional reduction, computing complexity and multi-resolution data approximation is being done. The ORL database includes 40 individuals' faces, each with 10 images (112×92). According to the experimental result, the recognition efficiency was improved and the speed rate of the recognition operation was increased by 86%.