Estimation and Prediction of Evolving Color Distributions for Skin Segmentation under Varying Illumination
Proceedings IEEE Conference on Computer Vision and Pattern Recognition. CVPR 2000 (Cat. No.PR00662) (2000)
Hilton Head, South Carolina
June 13, 2000 to June 15, 2000
Leonid Sigal , Boston University
Stan Sclaroff , Boston University
Vassilis Athitsos , Boston University
A novel approach for real-time skin segmentation in video sequences is described. The approach enables reliable skin segmentation despite wide variation in illumination during tracking. An explicit second order Markov model is used to predict evolution of the skin color (HSV) histogram over time. Histograms are dynamically updated based on feedback from the current segmentation and based on predictions of the Markov model. Translation, scaling and rotation in color space parameterize the evolution of the skin color distribution at each frame. Warping and re-sampling the histogram propagate consequent changes in geometric parameterization of the distribution. The parameters of the discrete-time dynamic Markov model are estimated using Maximum Likelihood Estimation, and evolve over time. Quantitative evaluation of the method was conducted on labeled ground-truth video sequences taken from popular movies.
L. Sigal, S. Sclaroff and V. Athitsos, "Estimation and Prediction of Evolving Color Distributions for Skin Segmentation under Varying Illumination," Proceedings IEEE Conference on Computer Vision and Pattern Recognition. CVPR 2000 (Cat. No.PR00662)(CVPR), Hilton Head, South Carolina, 2000, pp. 2152.