Face Recognition Based Principal Component Analysis and Wavelet Sub bands

Abstract

Face recognition is important in human identification. The biological recognition technique acts as a good method and has broad application in security areas. This work presents a method to improve the face recognition accuracy using a combination of Principal Component Analysis (PCA) and Wavelet Transform. Wavelet Transform is used to decompose the input image with different levels and rearrangement of a subband of the wavelet in a way that extracts good information from the image; PCA is used as data redundancy and takes the better representation of input data. We apply the proposed method to the standard face recognition dataset, the ORL data, and the dataset from our environment to make the proposed method be practical. The comparison for different levels of wavelet shows that the third level has better recognition accuracy with respect to other levels. Finally, the performance of the proposed method is compared with other methods and gives better recognition accuracy.