Fiding mean and variance of sample of gray scale images

Hi,

I am trying to find the mean and variance of 64 samples of images between each same pixel location of my image database. Each image is 192x164 resolution which I call from a directory to read in a cell variable 'data'. Then using for loop, I read and conver each image into column vector, double and store in variable 'M2' so it's dimension is 32256x64. Each column consist of pixel of each image sample. Now when I try to find mean along rows by applying mean2=mean(M2,2); matlab is giving this error Index in position 1 is invalid. Array indices must be positive integers or logical values.

A = dir('Input Folder'); 
TotalFiles = length(A);
data = cell(1, TotalFiles);
for k = 1:TotalFiles 
  data{k} = imread(A(k).name); 
end

% first=cell2mat(data([1]));
%  [row col]=size(first);
% M=double(zeros(Size(1),Size(2)));
% M1=M(:); Size=size(first);


 for i=1:TotalFiles
 M=double(cell2mat(data(i)));
%  M1(1:row*col,1)=M(:)+M1;
 M2(:,i)=M(:);
 end
 
% mean=(1/TotalFiles)*M1;
% mean1=reshape(mean,[192,168]);
mean2= mean(M2,2);

I don't understand why it is giving this error. Same way if I tried to find variance of this matrix, var2=var(M2,2), it gives same error. Kindly someone help me what is that I am doing wrong. I am using doule class for integers, Thanks

[Salman Imtiaz]

For mean, I am expecting mean2 matrix to be 32256x1 with each row location showing mean of 64 samples accross the data.

Answer 1

[Image Analyst]

Hmmm.. Do I tell you how to fix messed up code, or just tell you to not use messed us code in the first place? I think I'll do the latter. Don't use cells - no need for that complication. If you have enough memory and want to find the mean and standard deviation image, just read them in as slices in a 3-D array.

By the way, mean2 is a built-in function so you won't want to assign anything to mean2 like you did.

Something like (untested)

TotalFiles = length(A);
array3d = double(192, 164, TotalFiles);
for k = 1:TotalFiles 
    fullFileName = fullfile(pwd, A(k).name);
    array3d(:, :, k) = imread(fullFileName); % The image had better be grayscale!
end
meanImage = mean(array3d, [], 3);
stdImage = std(array3D, 0, 3)

[Salman Imtiaz]

Hi Image Analyst,

Thanks for the response, it is fixed. Yes images are grayscale and using cell as I need to show it. I can use array also as you mentioned. Also I didn't knew mean2 is built in function, did not find it on mathworks.

I would like to know if I use meanImage=mean(array3d,2), will it work same? what does "mean" syntax in your code signifies, array3d is variable with stored images, [ ] (all rows and columns?, and 3?

Thanks

[Walter Roberson]

mean2 is mean over the entire array . That is not the same as mean with dimension 2 which calculates the mean per row.

Answer 2

[Walter Roberson]

you accidentally defined a variable named mean and it is still in your workspace when you try to use mean as a function .

[Salman Imtiaz]

yes! I cleared my workspace before running the code and commented mean veriable. I knew I messed up using mean as variable when it is function. Thanks