Skin Color Segmentation based Face Detection Using Canny Edge Detection Method

Abstract: Skin segmentation aims to locate skin regions in an input image. An algorithm for segmenting skin regions and non-skin regions in color images using color space and edge information has been used here for face detection. We propose a face detection method based upon the edge and skin tone information of the input color image.  A pre-processing method uses color segmentation to achieve high performance in face detection. First, it consists of skin color information collection and segmentation. After skin color detection next step is to finding face among skin color detected region by using the golden ratio of face i.e. ratio of width and height of the human face. Here the first color image from input color space to RGB color space and then transferred into YCbCr and after this transformation, we have applied edge detection method to separate skin region and non-skin region. Here we have used two edge detection operators Canny and Sobel to find which operators detect better edges and we found that the performance of canny edge detection operator is much better than Sobel.

Keywords: Skin Segmentation, Canny Edge and Sobel Edge Detection, Color Model, the Golden Ratio of Face=1.5