I'm looking for these formulations entropy methods and also details on putting formulations in matlab for implementation of rough, tsallis, renyi entropy methods. Also, can you please recommend a book or any literature?
Dear S. Khattak,
In order to find out the probability of a gray scale image, first you have to convert the image into Integer format (uint8(img), where 'img' is input image). Then find out the histogram of that particular image (For dynamic range of pixel 256 you will get number of repetition of pixels from 0 to 255.). Probability of each pixel is nothing but number of repetition of each pixels divided by total size of image (i.e if h = imhist(uint8(img)), then probability P = h/(M*N) where M and N are row and column size of image). Once you find out the probability then you can apply the following formulas for entropy calculation ...
follow the links for entropy information
Shanon's Entropy = http://en.wikipedia.org/wiki/Entropy_(information_theory)
Renyié Entropy = http://en.wikipedia.org/wiki/R%C3%A9nyi_entropy
and You have to fix a q value for Renyi's Entropy other than q=0, Choose q=2 for best effect.
Regards,
Susant
Thank you, some papers refers to alpha in the formulations while some q are they same or does it make any difference?
@S. Khattak,
Alpha and q are same, it is known as the order of the Renyi's Entropy. You can think Shannon's entropy is a special case of Renyi's entropy i.e. if the order of the entropy (q or alpha) tends to zero then Renyi's entropy will converges towards Shannon's entropy.
- Badges
- Science topic