Adaptive propagation-based skin segmentation method for color images

Abstract

Segmentation of skin regions from color images has many important Computer Vision applications. But, accuracy of existing skin detection methods are severely affected by the color similarity between the background and actual skin regions. Probabilistic approaches using skin probability maps (SPMs) can solve this problem to an extent. In this paper, a novel method has been proposed which uses seeded region growing method. Region growing is implemented by an adaptive cost propagation and neighborhood analysis scheme. The initial seeds are obtained from the SPM. It is observed from the experimental results that the proposed method can perform better compared to the existing skin segmentation methods for different illumination and background conditions.