Image processing: finding transfer function of histogram equalization

Hello, lets say I have an image then I adjusted the contrast by using histogram equalization (histeq) Is there a way to plot or get an image of the transfer function that this command uses ?

Answer 1

[Image Analyst]

Just take the histogram of the original image, then call cumsum on the pixel counts and plot it.

[pixelCounts, grayLevels] = imhist(grayImage);
cdf = cumsum(pixelCounts); % Make transfer function (look up table).
cdf = cdf / sum(cdf); % Normalize
plot(grayLevels, cdf, 'b-');

[John Snow]

This is good thankyou very much

[Royi Avital]

This will work for Equalization. What if I want any other CDF function?

[Image Analyst]

Huh? That has nothing to do with equalization, though equalization can use the cdf just like a lot of other functions do. This is just general purpose code that will compute the cdf regardless of your histogram. What do you mean by "other CDF"? What other CDF functions might there be?

Answer 2

[Anand]

Or use the two argument syntax for histeq.

[histIm,T] = histeq(grayImage);
plot(T,'b-')

[Image Analyst]

I didn't know it had that output - I guess I should have looked. Of course like I warn people histogram equalization gives a nonlinear stretch which usually looks pretty crummy and unnatural. imadjust() almost always gives a better looking image.

[John Snow]

The result isnt "smooth" so I will take the first answer but thanks anyway