Skin Detection Technology Research And Improvement

Skin Detection Technology (Skin Detection) is a hot research topic involved with optical information processing, digital image processing, computer vision and pattern analysis and recognition. Skin detection plays a key role in many computer vision tasks like face detection and recognition, expression recognition, gestures recognition, human detection, content-based image and video indexing, etc, especially in adult image detection. It has been widely used in the area such as human machine interface, access control, surveillance and objectionable Web images filtering system and so on.In recent years, many studies have been done on human skin analysis and detection, and many useful conclusions have been obtained. This laid a sound foundation for researchers to understand the properties of human skin colors. This dissertation focuses on detection of skin color in static images.From the point of classification, skin detection and identification technology is a binary classification issues, it means a kind of method of mark up skin and non-skin pixels from an image. An outstanding skin color classifier's establishment needs to consider the color space selection and the skin color modeling technology and so on. And be able to adapt to the changes of all kind light conditions, the different characteristics of imaging equipment and a widely range of differences in skin color samples, and individual differences.In this thesis, some key problems on skin color detection are analyzed and studied soundly. At the moment, many skin detection methods only considered of the pixel-level color information, do not using other information of image in space, like texture information, and the impact of the lighting conditions. Based on SPM method, A new skin detection method using color and texture information is proposed.The main contributions of the thesis include:①Classify and summarize the existing research methods, discussed the common color space used for skin detection, the advantages and disadvantages of each color space, and is there an optimal color space for skin detection? discuss the existing skin color modeling technology and their respective characteristics.②Using CIE Lab color space histogram to detecting the color cast, if the image has no cast, then do not correct the color of the image.③Using GrayWorld algorithms to correct the color of image caused by non-standard illuminants and this can improve the performance of true detection.④Extract the texture information by using the co-occurrence matrix of gray image to filter pixels that the colors are skin-like.