How can i calculate the entropy of the region of interest of an RGB image?

I know how to calculate the entropy of an RGB image.First ,vectorize the image and finally use it as a parameter of the entropy() function.

    image_vector = double(original_RGB_image(:));

    entropy_of_the_image = entropy(image_vector);

But what should we do to calculate the entropy of the region of interest of image?

I know one approach.Somehow,convert it into a black-white image(white=1) and use the inverse of the black-white image as a parameter of the original RGB image.I think,then the background color will be black .But it doesnot work.The background(sourrounding) became blue(something like that) instead of black? what should i do? Here is my code...

    a=imread('ISIC_0000001 w.jpg');
    a=imresize(a,[300 300]);
    mkr=a;
    figure,
    imshow(mkr);
    im=rgb2gray(a);
    level=graythresh(im);
    a=im2bw(im,level);
    figure,
    imshow(a);
    mkr(a)=nan;
    figure,
    imshow(mkr);

Answer 1

[Image Analyst]

You can do this:

entropyWithinMask = entropy(im(a));

You might also want to look at entropyfilt() for a spatial filter based on entropy.

[Sadia Iffat]

ok.what does entropy of an image mean? I think on image purpose it measures the variance of colors.So,the number of different(most) colors present in an images means it has higher entropy then others images with a few colors. Am i right?

[Walter Roberson]

https://www.mathworks.com/help/images/ref/entropy.html

"Entropy is defined as

 -sum(p.*log2(p))

where p contains the histogram counts returned from imhist."

This does not imply that an image with more colors has higher entropy: it depends on how the bins are distributed.

[Sadia Iffat]

So,why entropy is used for image classification? How can i understand when i should measure entropy for the classification problem?