The influence of illumination on facial feature extraction

Abstract

Basically, the computer-based face recognition area is very challenging, especially when it comes to recognizing objects and images under different illumination conditions. One effective way to solve this problem is to examine the extent to which features remain available for extraction in differing illumination conditions. This paper studies the effect of different illumination conditions on feature extraction in improving the recognition of images. The Discrete Wavelet Transformation technique is used for feature extraction under different types of illumination. Based on the results of our experiments on the collected data set, it is concluded that features remain the same i.e. they are not affected by illumination conditions. It is also concluded that when features are extracted, light only affects the smooth areas of the face.