A review and comparative study of skin segmentation techniques for handheld imaging devices

A number of well-known algorithms for skin segmentation in unconstrained acquisition conditions are tested across many of the best known color spaces and some less well known ones. Although in theory each color space provides the exact same information, our results show that certain color spaces are more optimal where fast segmentation with limited resources is required. A comparative analysis and discussion of the results are presented and conclusions drawn as to the effectiveness of a range of different color spaces.