Optimal Combination of Color Components for Skin Detection

This paper proposes an optimal combination of color components that best represent skin across different ethnics. The optimal combination is selected based on the performance of various combinations of color components, namely a combination of a, b and g, a combination of U, V and g, a combination of H, r, g and b, a combination of Hnew, Snew and g, and a combination of Hnew, Snew, H and g, when incorporated with RGB and morphological filters. The methods used to detect the skin are based on threshold-based method. Experimented on three public datasets under various imaging conditions across different ethnics, we found that the combination of H, Hnew, Snew and g outperformed the other combinations.