what do you exactly need? Because these entropy formulations are pretty straight forward to implement in matlab. The probability distribution is, of course, the tricky part.. Once you have the distribution p, for example, for Shannon, you could do:
H = sum(-(p(p>0).*(log2(p(p>0)))));
You can easily modified the above to meet Tsallis and the others.
Thank you.How can I get probability for an image dataset to put in to different entropy formulations.
Also can you please recommend a book or any literature to look for these formulations entropy methods and details.
Hi. About the probabilities it could be as simple of count how many times a particular color, at pixel level, repeats in the whole image. So, for example, if you have an image with a size N = 10x10, with only 3 different colors 1, 2 and 3, in order to find the probability of a pixel to be 1, you only count how many 1's you have in the entire image (let say n) and then divided by N. So, P(1) = N/n. In matlab there's the function hist(X), where X is the image. This function calculates all the probabilities for you. About the literature I don't recall right now a single book that covers all types of entropy, particularly one that mix additive (Shannon) and non-additive (Tsallils) entropies. Maybe the best if look for them one by one in different sources, particularly in their source articles.
Thank you, some papers refers to alpha in the formulations while some q are they same or does it make any difference?
- Badges
- Science topic