Research Of The Skin Detection In The Complex Illumination

Skin Color Detection Technology is a hot research topic involved with the image recognition and understanding, it is the early stage of a lot of applications based on skin color processing, such as detection, gesture recognition, sensitive image recognition and filtering. Robust skin color detection technology is the key step for the success of such computer vision applications and systems. Although there have been many skin detection technology, and obtained a certain result, these are not practical in most applications, the complex skin detection technology under the complicated illumination is not mature.Therefore, this paper focuses on the study of skin detection technology under the complex illumination. It firstly introduces the basic theoretical knowledge of skin color detection, and then does the research of some relevant key techniques, completes the following major tasks:1. An improved Gamma correction method is proposed. This method use nonlinear methods instead of linear method to modify the Gamma values. This can enlarge the interval of Gamma values effectively and make the Gamma correction method more suitable to the changes of highlights, transition regions and shadows in images. So the proposed Gamma correction method is improved to have a more adaptability in illumination changes, and can weaken illumination affects on images effectively.2. Analysis of several common color space and color model, compare and the select the YCbCr space and elliptical skin model to carry on the preliminary skin detection. The YCbCr space is the color space of brightness-chromaticity separation space, which has the linear transform relations with RGB and its calculation is simple and rapid, while it has a greater advantage in video compression. The elliptical skin model and Gaussian model are belong to the same parameter model of the statistical model, the former is better than the latter in the test results on the processing speed and results, this paper experimentally analyses and compares the features and advantages and disadvantages of both. 3. We study the GLCM method to extract the image texture information, which use the direction and distance parameters between pixels to build co-occurrence matrix, and then extract meaningful statistics from the matrix, such as Energy, Contrast etc, to indicate texture feature. According to the texture information of skin samples, we build the texture characteristic function to determine skin areas and form the skin texture image.4. An adaptive algorithm under the complex illumination is proposed for skin detection, including the integrated illumination process, skin color and texture feature extraction, this algorithm is based on analysis and research in the small samples, which use the color and texture information to separate the skin areas. Experimental results show that the algorithm's accuracy and feasibility.