Color Constancy Using KL-Divergence

Color is a useful feature for machine vision tasks. However; its effectiveness is often limited by the fact that the measured pixel values in a scene are inJiuenced by both object surface reJiectance properties and incident illumination. Color constancy algorithms attempt to compute color features which are invariant of the incident illumination by estimating the parameters of the global scene illumination and factoring out its effect. A number of recently developed algorithms utilize statistical methods to estimate the maximum likelihood values of the illumination parameters. This paper details the use of KL-divergence as a means of selecting estimated illumination parameter values. We provide experimental results demonstrating the usefulness of the KL-divergence technique for accurately estimating the global illumination parameters of real world images.

[1]  Guillermo Sapiro,et al.  Color and Illuminant Voting , 1999, IEEE Trans. Pattern Anal. Mach. Intell..

[2]  L. Maloney,et al.  Color constancy: a method for recovering surface spectral reflectance , 1987 .

[3]  Ingemar J. Cox,et al.  An Analysis of Camera Noise , 1992, IEEE Trans. Pattern Anal. Mach. Intell..

[4]  Takeo Kanade,et al.  Rotation Invariant Neural Network-Based Face Detection , 1998, Proceedings. 1998 IEEE Computer Society Conference on Computer Vision and Pattern Recognition (Cat. No.98CB36231).

[5]  William T. Freeman,et al.  Bayesian decision theory, the maximum local mass estimate, and color constancy , 1995, Proceedings of IEEE International Conference on Computer Vision.

[6]  Graham D. Finlayson,et al.  Color by Correlation , 1997, Color Imaging Conference.

[7]  Graham D. Finlayson,et al.  Colour constancy with error bars , 1999 .

[8]  John K. Tsotsos,et al.  From [R, G, B] to Surface Reflectance: Computing Color Constant Descriptors in Images , 1987, IJCAI.

[9]  Heekuck Oh,et al.  Neural Networks for Pattern Recognition , 1993, Adv. Comput..

[10]  Brian V. Funt,et al.  Learning Color Constancy , 1996, CIC.

[11]  D H Brainard,et al.  Bayesian color constancy. , 1997, Journal of the Optical Society of America. A, Optics, image science, and vision.

[12]  Brian V. Funt,et al.  Color constancy under varying illumination , 1995, Proceedings of IEEE International Conference on Computer Vision.

[13]  Mark S. Drew,et al.  Diagonal transforms suffice for color constancy , 1993, 1993 (4th) International Conference on Computer Vision.

[14]  Brian V. Funt,et al.  Bootstrapping color constancy , 1999, Electronic Imaging.

[15]  Graham D. Finlayson,et al.  Colour by correlation: a simple, unifying approach to colour constancy , 1999, Proceedings of the Seventh IEEE International Conference on Computer Vision.

[16]  Kobus Barnard,et al.  Practical colour constancy , 1999 .