Hi,

3x3 pixels of an image means you want to compute variance for a 3x3 matrix. You can do so very simply:

Let's suppose A is that 3x3 matrix, then:

variance = var(A(:));

This should give you a single scalar value for variance.

Similarly, you can use std() for calculating standard deviation:

stdev = std(A(:));

Hope it helps.

Check this

http://angeljohnsy.blogspot.com/2011/09/mean-median-variance-standard-deviation.html

Let me leave some comment. From a theoretical point of view It seems to be not entirely reliable to calculate variance of 9 elements. Because the sample is very small and calculated estimation of the variance poorly reflects with the real value. For better results sample should contain 30-40 elements, 100 elements is much better.

You can use the blockproc Matlab built-in function. Just pass the image, block size (3x3 for your case), and the function (std or var).

In order to calculate the variance of matrix A, you can use:

Var = std2(A)^2; 

But, if you want to calculate local variance in patches of 3*3 pixels of image I, you can us the following command:

J = stdfilt(I);

Good luck

Shafaq Nisar

for variance, var(Image)

for standard deviation, std2(Image)

Hi to all,

Is there a tool in Python that can help visualizing the variance of images provided a bunch of images ? Let's say 4000 images. I want to now if there are very similar which I think would result in a poor dataset or if they are well different (provided the same class) which would infer that the images are well diverse in order to improve the training on image classes.

Thanks.