Illumination Correction for Static Skin Filters
Illumination correction or Color Constancy (CC) is the ability of humans for resolving the apparent colors of objects in a given scene independently of the illumination source. For robust skin detection in images/videos, we evaluate and investigate the effect of CC approaches on static skin segmentation approaches. The effect of five CC approaches namely Gray-Edge CC, Gray-World CC, max-RGB CC, Shades-of-Gray CC and Bayesian CC is studied for the YCbCr static skin filter. Results on two datasets (DS1 and DS2) having annotated per frame ground truth show that skin classification performance is improved. At the same time, the sift and deviation of the skin locus in the chromaticity space will result in alternating performance.
Keywords: Color constancy, skin detection, static skin detection
Download Full-Text
ABOUT THE AUTHORS
Rehanullah Khan
Sarhad University of Science and IT, Peshawar, Pakistan
Allan Hanbury
Institute of Software Technology and Interactive Systems, TU-Wien, Vienna, Austria
Amjad Khattak
UET, Peshawar, Pakistan
Farman Ali Khan
COMSATS institute of IT, Pakistan
Imran Maqsood
UET, Mardan, Pakistan
Rehanullah Khan
Sarhad University of Science and IT, Peshawar, Pakistan
Allan Hanbury
Institute of Software Technology and Interactive Systems, TU-Wien, Vienna, Austria
Amjad Khattak
UET, Peshawar, Pakistan
Farman Ali Khan
COMSATS institute of IT, Pakistan
Imran Maqsood
UET, Mardan, Pakistan