MATLAB: How to use built-in entropy function

digital image processingImage Processing Toolbox

Hi,

i would like to calculate entropy of an image using entropy function. when use this function like this

E = entropy(uint8(output_image)

i can see a value. But when i remove uint8, i see 0 as a result. type of output_image is double. Also when i try to calculate entropy like described in this link

E =  -sum(histogram_of_output_image.*log2(histogram_of_output_image));

i see NAN (not a number) as a result

Best Answer

Hi, Can you attach your "output_image" ? So that we can reproduce the issue.

Otherwise you can try the code below.

E = entropy(uint8(rescale(output_image, 0, 255)))

Related Solutions

MATLAB: Image processing using entropy

There is an entropyfilt() function in the Image Processing Toolbox.

MATLAB: Does the Image Processing Toolbox (R2007a) ENTROPY command return a value of zero when the image is cast as a double

Some of the Image Processing Toolbox functions have to assume a particular dynamic range. By convention, this range is [0,1] for data of type 'double', but [0,255] for data of type 'uint8'. Therefore, the IM2DOUBLE function should be used as it handles this scaling automatically. For example, your code thus becomes:

I = rgb2gray(ColorImage);           % use any true color image
ent1 = entropy(I);
I = im2double(I);
ent2 = entropy(I);

Both 'ent1' and 'ent2' should be equal.

Best Answer

Related Solutions

MATLAB: Image processing using entropy

MATLAB: Does the Image Processing Toolbox (R2007a) ENTROPY command return a value of zero when the image is cast as a double

Related Question