Human skin detection using defined skin region

Abstract

Skin detection has been widely used in image processing application. Several skin detection methods have been introduced including defined skin region. Overall, skin detection has been challenged by the fact that skin appearance in the image can be affected by many factors including ethnicity thus reducing its accuracy. This study aims (i) to obtain the skin detection model by ascertaining the threshold values of human skin color model and (ii) to evaluate the skin detection model among ethnicities and to evaluate the accuracy of the model by comparing the model in human and animal skin. The result of our study show that the value of skin parameter is not different among ethnicities. We also found that using this method, animal skin was often misinterpreted as human skin. This problem should be particularly addressed in the future work