Study On Face Detection In Complex Background

Recognition and analysis of human faces can be widely used in such fields as personal identification recognition, safety inspection, human-computer interaction, expression analysis, lip-reading, etc. As the premise and basis of human face recognition and analysis, human face detection attracted people's attention at a very early time with the help of computer. Along with the generalization of computer application, the improvement of computer performance and achievements in image processing and pattern recognition fields, human-face concerned applications is getting nearer and nearer to practical use, thus resulting in more and more emphasis over research for human face detection.The front face detection of color still images is discussed in this paper. The problems in the existing face detection methods based on the research for the skin-tone-based detecting algorism are analyzed. It is discovered that the successful algorisms for human-face detection aim at the images in simple background, however, as to images in complex background, existing algorisms either are slow at detecting or easily lead to detection failure, and the effect is not very well.The YCbCr color space is chosen to segment the skin tone. In the space, the region of the skin color is established by nonlinear method. And it can get the polyhedron in the color space. The parameter of the fuzzy function is gotten by repetitious tests. By combining the two spaces, a new method of segmenting skin tone, the method based on dual matching of skin color and region, is proposed in this paper. The new method not only based on the information of skin tone, but also the information of region, which is easily ignored in the existing algorism. The new method increases the relativity of the skin-tone and region. Then the noise of the picture is wiped off by using morphological approach, which gets a good result and locates the candidate human face effectively.Moreover, the following method is used to locate the position of the eyes in this paper. First, two separate eye maps are built according to the value of Y, Cb and Cr,...