Image processing techniques may be used for enhancing edges, boundaries, contrast, etc. of an image through accentuation or sharpening process. Many algorithms are available to normalize illumination impact for different image processing applications. In this contribution, we conduct a comparative study on four different types of illumination normalization algorithms. They are based on discrete wavelet, logarithmic total variation with primal dual algorithm, histogram equalization technique, and morphological operation. In order to obtain better enhancement of image by using discrete wavelet based illumination normalization algorithm, selection of the particular wavelet is very important. Use of histogram equalization function in image preprocessing with other algorithms enhances the overall performance of face detection. This paper illustrates performance of different illumination normalization algorithms in Viola-Jones face detection system based on the extended Yale B database.