The Student Room Group

Histogram - bins values?

I have come across a definition of a term, uniformity which is
"the sum of the square of the histogram bins value".

I don't understand what this means. Please can someone explain what it means? thanks
Original post by jsmith6131
I have come across a definition of a term, uniformity which is
"the sum of the square of the histogram bins value".

I don't understand what this means. Please can someone explain what it means? thanks


Judging from the first few paragraphs of this wikipedia article I would interpret a 'bin value' as the number of things in the 'bin', which are the intervals, and you square that and sum it up. If this is misleading I shall delete my post.

Edit: seems to be misleading.
(edited 8 years ago)
Original post by SeanFM
Judging from the first few paragraphs of this wikipedia article I would interpret a 'bin value' as the number of things in the 'bin', which are the intervals, and you square that and sum it up. If this is misleading I shall delete my post.


well that's what I thought it meant but it didn't make much sense to me...but if you agree then I suppose thats good enugh :smile:
Original post by jsmith6131
well that's what I thought it meant but it didn't make much sense to me...but if you agree then I suppose thats good enugh :smile:


Does it make sense for the defintion/context?
Original post by SeanFM
Does it make sense for the defintion/context?


well the context is uniformity in images. If you consider the image in terms of its pixel values then you get a histogram. An image with 100 pixels of the same value should have a uniformity of 1. So I don't know how the definition fits...unless it is considering the probability value squred?
could it be talking about doing a chi squared test (of uniformity)?

probably more context would help.
Original post by jsmith6131
well the context is uniformity in images. If you consider the image in terms of its pixel values then you get a histogram. An image with 100 pixels of the same value should have a uniformity of 1. So I don't know how the definition fits...unless it is considering the probability value squred?


Ah, I don't think what I said is correct then.


sorry, still non the wiser really. I'd guess he's talking about some sort of edge detection algorithm that he's chosen to call 'uniformity' ?

The text seems to have been translated into English quite crudely.

Quick Reply

Latest