A Power Law Transformation Predicting Lightness Conditions Based on Skin Color Space Detection

This paper presents an improved human-skin-color detection approach in images using power law transformation predicting lightness conditions. Our approach focus on different illumination conditions and complex background via RGB-HSV model. The illumination is adjusted without loosing image quality, and skin detection is done in a per pixel fashion after transforming into HSV color space. The experimental results demonstrated that, the proposed technique is able to achieve robust results, where a set of skin color images from ethnic color images from ethnic groups have been detected such as Asian, African and Caucasian.