I have a RGB image which i converted into HSV color model. Then I have extracted the V- component. I draw simple histogram of this V-component. Now i wanted to divide this created histogram in such 8 part where each partition should have the same ent

I have a RGB image which i converted into HSV color model. Then I have extracted the V- component. I draw simple histogram of this V-component. Now i wanted to divide this created histogram in such 8 part where each partition should have the same entropy. Or if not 8 then number of partition must be same.

[Fahad Mairaj]

would you plz share your code? this will be helpful for me.

[Im_possible_woman]

Hello @sumit kumar, could you please share the codes of RGB image which is converted into HSV color model. Then have extracted the V- component? I really hope that you can help me to share the codes and respond me.

Answer 1

[Image Analyst]

Compute the histogram. Then compute the entropy in each bin: sum(p*log(p)). Then compute the CDF of the entropy array using cumsum(). Then split it at the 1/8, 2/8, 3/8, etc. points to find out what gray levels give you the eight part splitting that you seek.

[sumit kumar]

Thank you sir. One more thing by entropy array do you mean that when i am calculating entropy for each bin then each value must be stored in an array and that array is called entropy array.?????/

[sumit kumar]

can you elaborate. if i apply sum(p*log(p)) then i will get single entropy for entire image? Secondly how can i proceed to divide in 1/8 2/8 3/8...... please explain this a bit. Thank you

Answer 2

[Walter Roberson]

You could use quantile

[sumit kumar]

is it a in-build matlab command???

[Walter Roberson]

It is part of the Statistics and Machine Learning Toolbox

[Walter Roberson]

Though I am not sure what you mean by "have the same entropy" with respect to a portion of a histogram?