Skin detection using color distance measurement and thresholding

Abstract

Segmentation is a process to extract the objects in an image. In this study, segmentation process is used to perform skin detection which aims to detect skin pixels and areas in an image. Pixel-based measures is used in the similarity measurement model in image area segmentation to detect skin and non-skin areas in 3 color spaces i.e. RGB with Euclidean distance, HSV with Dcyl distance, HCL with DHCL distance and the thresholding value as the limit of the similarity between two colors. The results of this study indicate that color space segmentation process using thresholding value (Th 80 and Th 100) for HSV and HCL provide superior segmentations.