How can I get the histeq function to ignore NaN values?

I have a set of 2D pictures, I want to only focus on a small blob in the middle while ignoring a sea of NaNs around it

If I keep the NaNs, histeq spits out a plain white image. Is there a way to stop NaNs from corrupting the histeq function?

thanks in advance

- Nadeem

Answer 1

[Ramnarayan Krishnamurthy]

It appears that MATLAB warns you about using histeq on a matrix with NaNs and replaces them to 0.

I would suggest the following if you want to apply histeq only to non-NaN values in a matrix

% Setting up a sample image 
% Use im2double as double saves NaN while uint8 converts it to 0

I = im2double(imread('cameraman.tif'));

I(1:100,1:100) = NaN
I(1:256,1:100) = NaN
I(1:256,180:end) = NaN
I(1:100,:) = NaN
I(160:end,:) = NaN

figure, imshow(I)
% Duplicate Matrix
I2 = I;

% Store the output of the contrasted stretched image in I2
I2(~isnan(I2)) = histeq(I2(~isnan(I2)));
figure, imshow(I2)

Hope this helps!

Answer 2

[Image Analyst]

I wouldn't even use histeq(). I mean, WHY? It often/usually gives crummy, harsh, unnatural looking images, and it's not needed to do image analysis.