Skin segmentation based on multi pixel color clustering models
Sinan A. Naji,Roziati Zainuddin,Hamid A. Jalab
DOI: https://doi.org/10.1016/j.dsp.2012.05.004
IF: 2.92
2012-12-01
Digital Signal Processing
Abstract:This paper presents a reliable color pixel clustering model for skin segmentation under unconstrained scene conditions. The proposed model can overcome sensitivity to variations in lighting conditions and complex backgrounds. Our approach is based on building multi-skin color clustering models using the Hue, Saturation, and Value color space and multi-level segmentation. Skin regions are extracted using four skin color clustering models, namely, the standard-skin, shadow-skin, light-skin, and high-red-skin models. Moreover, skin color correction (skin lighting) at the shadow-skin layer is used to improve the detection rate. The experimental results from a large image data set demonstrate that the proposed clustering models could achieve a true positive rate of 96.5% and a false positive rate of approximately 0.765%. The experimental results show that the color pixel clustering model is more efficient than other approaches.
English Else