Improving Color Constancy by Photometric

Improving Color Constancy by Photometric

Edge Weighting

ABSTRACT:

Edge-based color constancy methods make use of image derivatives to estimate the illuminant. However, different edge types exist in real-world images such as material, shadow and highlight edges. These different edge types may have a distinctive influence on the performance of the illuminant estimation.

Therefore, in this paper, an extensive analysis is provided of different edge types on the performance of edge-based color constancy methods. First, an edge-based taxonomy is presented classifying edge types based on their photometric properties. Then, a performance evaluation of edge-based color constancy is provided using these different edge types. From this performance evaluation it is derived that specular and shadow edge types are more valuable than material edges for the estimation of the illuminant. To this end, the (iterative) weighted Grey-Edge algorithm is proposed in which these edge types are more emphasized for the estimation of the illuminant.

Images that are recorded under controlled circumstances demonstrate that the proposed iterative weighted Grey-Edge algorithm based on highlights reduces the median angular error with approximately 25%. In an uncontrolled environment, improvements in angular error up to 11% are obtained with respect to regular edge-based color constancy.

Architecture:

EXISTING SYSTEM:

The disadvantage of the proposedmethod is that in case of misclassification of the edgetypes the method may result in lesser performance.The weighted Grey-Edge inherits the weakness of the

regular Grey-Edge: opaque parameter selection, i.e.the optimal parameters are difficult to select withoutprior knowledge of the input images.

PROPOSED SYSTEM:

We propose a weighted Grey-Edge over existing methods is thatadditional information, which is provided by the differentedge types, is used. This information, whenavailable, results in more accurate illuminant estimates.Moreover, the proposed method connects twotheories involving color image analysis. The Grey-Edge for color constancy and the Quasi-Invariants for

Edge classification. The disadvantage of the proposedmethod is that in case of misclassification of the edgetypes the method may result in lesser performance.The weighted Grey-Edge inherits the weakness of theregular Grey-Edge: opaque parameter selection.The optimal parameters are difficult to select withoutprior knowledge of the input images.

MODULES:

1. Color Constancy.

2. Edge Classification.

3. Grey-Edge

4. Edge-based Color Constancy

Module Description

Color Constancy:

Extending pixel-based methods to incorporate derivative information, edges and higher-order statistics, resulted in the Grey-Edge and the derivative based Gamut mapping algorithms.The Grey-Edge actually comprises a frameworkthat incorporates zeroth-order methods, first-order methods, as well as higher-ordermethods .Many differentalgorithms can be created by varying the three parametersanother pixel-based method which has been extendedto incorporate derivative information is theGamut mapping algorithm. It can be proventhat linear combinations of image values also formgamuts, thereby extending the Gamut mapping theoryto incorporate image derivatives. In this paper,we assess the influence of various edge types on the performance of both the Grey-Edge method and thederivative-based Gamut mapping method.

Edge Classification:

Various edge types are considered, i.e. material edges, shadow or shading edges, specular edgesand interreflection edges. Material edges are transitionsbetween two different surfaces or objects. Shadingedges are transitions that are caused by the geometryof an object, for instance by a change in surfaceorientation with respect to the illumination. Shadowedges are cast shadows, caused by an object thatblocks the light source. Blocking of the lightsource often results in merely an intensity gradient,but sometimes a faint color gradient is introduced.When we refer to shadow edges in general, both intensityand colored shadow edges are implied. Finally,in real-world images, interreflection are an importantaspect. Interreflection is the effect of light reflectedfrom one surface onto a second surface. This effectchanges the overall illumination that is received bythe second surface, and hence the color of this surface.Finally, note that combinations of edge types can also occur, but are not handled explicitlyhere.

Grey-Edge:

It can be derived that thespecular and shadow-shading-specular variants andquasi-invariants are dependent on the color of thelight source .The underlying assumption thatthe scene is viewed under a white light source is obviously not met for the images inthe used data sets prior to applying color constancy.However, after the proposed algorithm is applied,the illuminant should be neutral, at least in theory.Hence, we propose to first correct the input imageA with an estimated illuminant I. Then, using thiscolor corrected image B, we can compute a weightingscheme W, which in turn is used by the weighted Grey-Edge algorithm to compute an updated estimate ofthe illuminant I. After some iterations, the illuminantestimate will approximate a white light source, atwhich point the accuracy will no longer increaseand the method has converged. Consequently, wepropose to iteratively apply the weighted Grey-Edgealgorithm, where a new instantiation of the weightingscheme is computed every iteration based on thecolor corrected image at each iteration.

Edge-based Color Constancy:

Extending pixel-based methods to incorporate derivativeinformation, i.e. edges and higher-order statistics,resulted in the Grey-Edge [14] and the derivative based Gamut mapping algorithms.The Grey-Edge actually comprises a frameworkthat incorporates zeroth-order methods (e.g. the Grey-World and the White-Patch algorithms), first-ordermethods, as well as higher-order methods.

Algorithm 1 Grey-Edge

Input:

input image: A

initial illuminant estimate: I

stopping criterion: C

Method:

while(: C) do

B = color_correct(A, I)

W = compute_weighting_scheme(B)

I’ = weighted_GreyEdge(B,W)

I = I _ I’

C = update_stopping_criterion()

end while

System Requirements:

Hardware Requirements:

System : Pentium IV 2.4 GHz.

Hard Disk : 40 GB.

Floppy Drive: 1.44 Mb.

Monitor: 15 VGA Colour.

Mouse: Logitech.

Ram: 512 Mb.

Software Requirements:

Operating system : Windows XP.

Coding Language: ASP.Net with C#

Data Base : SQL Server 2005