How would I create color commonality Image?

I am working on my research and I have a confusion regarding 'Color commonality Image'.

This type of image is used to address the problems in contrast in underwater images. It enhances the uncommon colors as suggested in paper.Basically, this type of image is obtained by replacing each pixel value of the illumination corrected image with the likelihood of the color across the entire image.

To derive such image, 3D color histogram is created(whether 64 bins or 256 bins in each dimension) from an RGB image(illumination corrected). This histogram is smoothed via 3D gaussian. Pixel value Icc(X) is entry Hrgb(Irgb(X)) in this histogram. It is applied to each pixel separately.

I have a confusion regarding the concept of the way "Color Commonality" Image. Exactly how this image is obtained.

I would appreciate if anyone tell me some other image transformations that are used to address the contrast and color information of RGB images(particularly underwater images).

Answer 1

[Image Analyst]

The description sounds exactly like the "Color Frequency Image". If it is, you can use this code to compute it: http://www.mathworks.com/matlabcentral/fileexchange/28164-color-frequency-image

[Jay]

Its the same thing!! Excellent. Thanks.