Separating Diffuse and Specular Reflection Components
based on Chromaticity and Noise Analysis


IEEE Transactions on Pattern Analysis and Machine Intelligence (PAMI), 26(10), pp. 1373-1379, October 2004

 


  Robby T. Tan      Ko Nishino *       Katsushi Ikeuchi

 


[ PAMI Paper ]

 

textured surfaces


Abstract:
Many algorithms in computer vision assume diffuse only reflections and deem specular reflections to be outliers. However, in the real world, the presence of specular reflections is inevitable, since there are many dielectric inhomogeneous objects which have both diffuse and specular reflections. To resolve this problem, we present a method to separate the two reflection components.  The method is principally based on the distribution of specular and diffuse points in a two-dimensional maximum chromaticity-intensity space. We found that, by utilizing the space and known illumination color, the problem of reflection component separation can be simplified into the problem of identifying diffuse maximum chromaticity. To be able to identify the diffuse maximum chromaticity correctly, an analysis of the noise is required, since most real images suffer from it. Unlike existing methods, the proposed method can separate the reflection components robustly for any kind of surface roughness.


Experimental Results:

"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.”


I. Evaluation:

1. Head Model

  

           (a) input image        (b) diffuse component (polarizing filters)   (c) estimated diffuse component

Comparison of (a) and (b):

  

Comparison of (b) and (c):

   

(d) error in R-channel                       (e) error in G-channel                      (f) error in B-channel

2.

  

(a) input image                (b) diffuse component (polarizing filters) (c) estimated diffuse component

Comparison of (a) and (b):

  

(d) difference in R-channel                    (e) difference  in G-channel                      (f) difference in B-channel

Comparison of (b) and (c):

  

(g) difference in R-channel                    (h) difference  in G-channel                      (i) difference in B-channel

 

III. Real Images:

1. Head Model (a single illumination)

   

   
 
2. Head model lit with multiple light sources:

   

3. A green plastic object with  rough surface:
   
   
  

4. Green sandal (high specularity):

   

   

5. Toy (low specularity)

   

 

[ abstract | experimental results | paper ]

 

 

 byRobby T. Tan
The Univesity of Tokyo