Image Enhancement Using Gamma and HE For the Face Detection System In Real-Time By Viola_Jones Algorithm

Abstract

The process of detecting and recognition faces in real-time play an important role in many systems. Face detection is the first stage of face recognition systems. The poor lighting conditions effect on the results, significantly, in the case of using the Viola algorithm to detect faces in real-time, so image processing is a very important step in improving the accuracy Viola algorithm. Most systems use a single technique that handles all images, without distinguishing between the images in which the viola can detect the face or not, which costs unnecessary effort to the system, In addition to, the difference in the type of damage to the image that needs to be process varies according to the type of problem. In this paper, a simple and fast method presented based on studying the histogram of the image that suffers from a problem and tends to be dark or bright. Then the system uses the gamma only, or it uses both gamma and histogram equalization (HE) to processing the image with variables that differ according to the type of problem that it caucused by the lighting of the image. In this study, three types of datasets are used. Each type contains images that are taken in different lighting conditions. The final results proved the increase in the ability of the Viola algorithm to detect faces in easy and the accuracy reached 95.8%.