A New Image Correction Method for Multiview Video System

Because of scene illumination or camera calibration, color appearance of the same object between different viewpoints may be different in multiview video system. Traditional illumination compensation algorithm for image is unable to solve this problem effectively. In this paper, a novel color correction method for multiview video system is proposed based on retinex color constancy theory. To eliminate influence of un-consistent light sources, histogram equalization, retinex processing and color restoration are performed for multiview images to extract reflectance that describes object intrinsic properties. Experimental results show that the proposed image correction method for multiview video system is effective

[1]  Yuichi Ohta,et al.  An approach to color constancy using multiple images , 1990, [1990] Proceedings Third International Conference on Computer Vision.

[2]  Toshiaki Fujii,et al.  Free viewpoint TV system based on ray-space representation , 2002, SPIE ITCom.

[3]  Michael Elad,et al.  A Variational Framework for Retinex , 2002, IS&T/SPIE Electronic Imaging.

[4]  E. Land.  The retinex theory of color vision. , 1977, Scientific American.

[5]  Marc Levoy,et al.  Light field rendering , 1996, SIGGRAPH.

[6]  Sang Hyun Kim,et al.  Fast local motion-compensation algorithm for video sequences with brightness variations , 2003, IEEE Trans. Circuits Syst. Video Technol..

[7]  Richard Szeliski,et al.  The lumigraph , 1996, SIGGRAPH.

[8]  Hiroshi Watanabe,et al.  Global brightness-variation compensation for video coding , 1998, IEEE Trans. Circuits Syst. Video Technol..