Adaptive skin-color filter

In this paper, we propose a new method, called an adaptive skin color-filte r, for detecting skin color regions in a color image. The use of skill colo r provides an efficient way to find candidate regions for faces or hands in color images. However, it is not easy to find skin color regions because t he color of skin regions varies from image to image due to a variety of rea sons. Since most of the previous methods adopt a fixed threshold scheme, th ey are useful only in the restricted (i.e., controlled) environment. Howeve r, our method is applicable to images in more general situations since it i s capable of adaptively adjusting its threshold values and effectively sepa rating skin color regions from similar background color legions. Our method consists of basically two stages. In the first stage, a thresholding box i n HSV color space is updated adaptively using a color histogram under the a ssumption that the area of skin color regions is comparable to that of simi lar background color ones. In the second stage, color vectors inside the th resholding box are classified into two groups: skin color vectors and backg round colors vectors. Our method and other conventional methods were tested with 379 images obtained from the Internet. Experimental results show that our method is robust to the variations of skin regions' color compared to the conventional methods. (C) 2001 Pattern Recognition Society. Published b y Elsevier Science Ltd. All rights reserved.