LEARNING SKIN DISTRIBUTION USING A SPARSE MAP (MA-P1)

Author(s) : Rajkiran Gottumukkal (Old Dominion University, USA) Vijayan Asari (Old Dominion Univeristy, USA)

Abstract : We present a new skin modeling technique based on SNoW (Sparse Network of Winnows) for accurate and robust skin region detection. A Skin Distribution Matrix (SDM) representing the sparse network is trained with skin pixels to learn the distribution of the skin pixels in a color space. We then train the SDM with non-skin pixels to forget the distribution of the non-skin pixels which overlap with the skin pixels in the color space. The accuracy of this skin model is tested over different color spaces. The performance of skin detection using SDM is compared with the Skin Probability Map (SPM) method. SDM needs lesser resources since the skin and non-skin distributions are represented by a single matrix to perform skin detection in color images.

Menu