Human skin detection in digital images using multi colour scheme system
Shanmugavadivu Pichai, Ashish Kumar · 2017
A new combinational and Multi-Colour Scheme System for the digital images is presented in this paper. This preprocessing algorithm finds application in human face detection as well as in recognition. The Multi-Colour Scheme System (MCSS) for Human Skin Detection in digital images aims at object localization, based on principle of neighborhood pixel processing. This technique due to its computational accuracy and performance is proved to have an edge over the competitive methods. It also assures pay-off on reduced computational complexity, compared to facial textures/ geometrics-based skin detection methods. A thorough comparative analysis is carried out to evaluate the effect and influence of different multi-colour bands for the human skin classification. MCSS primarily uses different combinations of multi-colour space namely, RGB (Red, Blue and Green), H (H component of HSV), CMYK (Cyan, Magenta, Yellow and Black) and a*b* (CIE-Lab, Commission internationale de l'éclairage). This multi-colour scheme based skin detection technique is tested on the real-time dataset and it is proved to produce more accurate results on skin detection than those competitive methods which employ single colour space. Due to its computational advantages, this method is confirmed to be superior to the recently reported competitive methods. The performance of MCSS is validated by Human Visual Perception.