Calculate SNR pixel-by-pixel

I am trying to figure out the Signal to Noise Ratio (SNR) in an image matrix (64x64x16). (called Image_Matrix_3D in the code)

The SNR = mean signal / stdev

However, I need to compute this pixel by pixel wise. so I need the mean signal in each pixel for each slice and the stdev in each pixel for each slice.

The code should look something like the following I think. However, it doesn't look entirely correct. Also I get the stdev to be zero.

for i=1:64
    for j=1:64
        for k=1:N
            m(i,j,k) = mean(Image_Matrix_3D( i,j, k),[3]);
            stdev(i,j,k)= std(m( i,j, k));
        end
    end
end

If anyone on Matlab Central is familiar with this type of analysis, can you please help?

Thanks!

Answer 1

[Walter Roberson]

The standard deviation of a single value is always 0.

You need be applying mean() and std() to the same vector of values. "vector" is a key term here.

Hint:

pixelvector = squeeze(Image_Matrix_3D(i,j,:));
mean(pixelvector)
std(pixelvector)