Hi! Question about image processing.

So I have here different retinal images and they all have different 'contrast'. What can I do to make these images have the same 'intensity' or 'contrast' or 'appearance' between them. Is there a way where for example an image seems dark, it will be contrasted to a lighter image and if it's a light image, it will be contrasted the other way around?

IMAGE 1: http://imgur.com/eXlrPfE IMAGE 2: http://imgur.com/TokFHka

Answer 1

[Image Analyst]

You can try histogram matching like this algorithm: http://www.eyemaginary.com/Portfolio/ColorHistogramWarp.html

The Image Processing Toolbox also has a imhistmatch() function in it. I also have a histogram matching routine in my File Exchange http://www.mathworks.com/matlabcentral/fileexchange/?term=authorid%3A31862 where you can make the histogram be any shape you like, for example below I make the histogram be the shape of the girl's silhouette:

Or you can play around with imadjust().

You might also look into HDR, High Dynamic Range Imaging.

[Portgas Ace]

hi! the link you've given about the colorhistwarp by mark grundland seems to be promising. But i cant seem to find a sample code for me to work with. We're allowed to use already working codes as long as we have understand how does it happen. If you could help me find a code to perform the colorhist warping, I would appreciate it really much.

[Image Analyst]

Sorry, I'd like to have the Grundland algorithm in MATLAB code but I don't. If you code it up, please put it on the File Exchange. It is the best color histogram matching algorithm I've seen.