Performance evaluation for face recognition using wavelet-based image de-noising

In this research we scrutinize the face recognition system performance when the test images are imposed to different levels of noise. We tried to imitate the real world scenarios when the face images are captured from video cameras or scanners and suffer some noise. To investigate the performance of proposed system, we simulate this scenario by adding A WGN (additive white Gaussian noise) to the test images in the face database. For image de-noising, we used two different algorithms namely; Discrete Wavelets Transform (DWT) and Dual-Tree Complex Wavelets Transform (DTCWT). The denoised images are then fed to a PCA-based face recognition system for better recognition performance.