Abstract:
In the past few years, there have been proposed many algorithms for face detection, using different approaches. Algorithms for detecting faces are used in the vision systems, robotics, video surveillance and access control systems. The problem of face detection has serious practical per-spective and interest for a great research. Often, it is the “first step” in the process of solving the problems of a higher level (for example, recognition of faces, facial expression recognition). For the successful system, face detection algorithms are necessary to provide high-speed operation and minimal number of false detection. The method of Viola–Jones is one of the best indicators for the effectiveness of recognition ratio/performance. However, this method in many cases gives a large number of false detection. The color of human skin is one of the features that helps make detection of face. This paper solved the problem of detection of face area by using the proposed method, which based on the combination of Viola & Jones method with skin detection method by using Log opponent and YIQ color spaces. To identify faces in the images, it is proposed to use, firstly, the Viola–Jones method for the original image. Then for each output region, the skin detection method is used to classify the evaluated area of the human skin. The experimental results show that the proposed method detected 95.75% of the faces presented on a set of test images, and substantially reducing the probability of the false detection.
Keywords:face detection, skin detection, Viola–Jones method, Log opponent color space, YIQ color space.
Fulltext:
PDF file (435 kB)