Illumination Chromaticity Estimation using
Inverse-intensity Chromaticity Space


IEEE Computer Society Conference on Computer Vision and Pattern Recognition 2003 (CVPR'03),
Madison,Wisconsin, USA. June 16-22, 2003

 


  Robby T. Tan      Ko Nishino *       Katsushi Ikeuchi

 



Journal of the Optical Society of America A (JOSA A):
"Color Constancy through Inverse-Intensity Chromaticity Space"
[PDF]


Abstract:

Existing color constancy methods cannot handle both uniform colored surfaces and highly textured surfaces in a single integrated framework. Statistics-based methods require many surface colors, and become error prone when there are only few surface colors. In contrast, dichromatic-based methods can successfully handle uniformly colored surfaces, but cannot be applied to highly textured surfaces since they require precise color segmentation.  In this paper, we present a single integrated method to estimate illumination chromaticity from single/multi-colored surfaces. Unlike the existing dichromatic-based methods, the proposed method requires only rough highlight regions, without segmenting the colors inside them. We show that, by analyzing highlights, a direct correlation between illumination chromaticity and image chromaticity can be obtained. This correlation is clearly described in ``inverse-intensity chromaticity space'', a new two-dimensional space we introduce. In addition, by utilizing the Hough transform and histogram analysis in this space, illumination chromaticity can be estimated robustly, even for a highly textured surface. Experimental results on real images show the effectiveness of the method.


 

"Do not use the images in this website for testing your code.

The images are compressed images whose brightness might not be linear to the flux of incoming light.”


1. Uniformly Colored Surfaces:
    (a). Head model illuminated by a halogen lamp:

headmodelsolux

    Estimation results: Er = 0.3779, Eg = 0.3242, Eb = 0.2866
    White Reference: Er = 0.3710, Eg = 0.31855, Eb = 0.31031

   (b) Head model illuminated by incandescent lamps:

headmodeltungsten

    Estimation results: Er = 0.4613, Eg = 0.3158, Eb = 0.2084
    White Reference: Er = 0.502146, Eg = 0.29838, Eb = 0.19946

    (c). Green sandal illuminated by a solux halogen covered by green filter:

sandalgreen
    Estimation results: Er = 0.3067, Eg = 0.4546, Eb = 0.2404
    White Reference: Er = 0.29804, Eg = 0.45807, Eb = 0.24387

   (d). Green sandal illuminated by a solux halogen covered by purple filter

sandalpurple
    Estimation results: Er = 0.3336, Eg = 0.0584, Eb = 0.5797
    White Reference: Er = 0.3336, Eg = 0.06483, Eb = 0.601529

2. Multicolored Surface:
    (a). Toy illuminated by incandescent lamps:

toytungsten
    Estimation results: Er = 0.5304, Eg = 0.2817, Eb = 0.2262
    White Reference: Er = 0.502146, Eg = 0.29838, Eb = 0.19946

    (b). Toy illuminated by a halogen lamp covered by blue filter:

toyblue
    Estimation results: Er = 0.2841, Eg = 0.3160, Eb = 0.3817
    White Reference: Er = 0.26335, Eg = 0.2986, Eb = 0.43801
   
    (c).
Toy illuminated by a halogen lamp covered by green filter:

    (c). Magazine illuminated by fluorescent lamp covered by green filter:

magazinegreen
    Estimation result: Er=0.2969, Eg=0.4877, Eb=0.2299
    White reference: Er=0.2828, Eg=0.48119, Eb=0.2359

    (d). Magazine illuminated by a halogen covererd by blue filter:

magazineblue
    Estimation results: Er = 0.2440, Eg = 0.3448, Eb = 0.4313
    White Reference: Er = 0.26335, Eg = 0.2986, Eb = 0.43801

    (d). Magazine illuminated by a halogen:

magazinesolux
    Estimation results: Er = 0.3368, Eg = 0.3285, Eb = 0.3410
    White Reference: Er = 0.3710, Eg = 0.31855, Eb = 0.31031

    (e) Complex scene illuminated by fluorescent light in uncontrolled environment

scene
    Estimation results: Er = 0.3210, Eg = 0.3465, Eb = 0.3093
    White Reference: Er = 0.33692, Eg = 0.34071, Eb = 0.31236

 

 by.  Robby T. Tan
The Univesity of Tokyo