Chapter

Computer Vision – ACCV 2007

Volume 4843 of the series Lecture Notes in Computer Science pp 728-737

Color Constancy Via Convex Kernel Optimization

  • Xiaotong YuanAffiliated withCenter for Biometrics and Security Research, Institute of Automation, Chinese Academy of Science, Beijing,100080
  • , Stan Z. LiAffiliated withCenter for Biometrics and Security Research, Institute of Automation, Chinese Academy of Science, Beijing,100080
  • , Ran HeAffiliated withCenter for Biometrics and Security Research, Institute of Automation, Chinese Academy of Science, Beijing,100080

* Final gross prices may vary according to local VAT.

Get Access

Abstract

This paper introduces a novel convex kernel based method for color constancy computation with explicit illuminant parameter estimation. A simple linear render model is adopted and the illuminants in a new scene that contains some of the color surfaces seen in the training image are sequentially estimated in a global optimization framework. The proposed method is fully data-driven and initialization invariant. Nonlinear color constancy can also be approximately solved in this kernel optimization framework with piecewise linear assumption. Extensive experiments on real-scene images validate the practical performance of our method.