finding the minimum value

I have a code performing inverse dualtree dwt

for j1 = 1:size(A1_20,3)

    for j = 1:J
    % loop thru subbands
    for s1 = 1:2
        for s2 = 1:3
            w{j}{s1}{s2} =A1_20(:,:,j1);
            y1 = idualtree2D(w, J, Fsf, sf);
            Er= mean2((single(x) - single(y1)).^2);
            final_col2{j1}=Er;

end
end
end
end

I have 5 images w denotes it,and A1_20(:,:,j1) contains 20 matrices of ,I have multiplied those matrices,with image so for each image i get 20 images and have found error,now i want to display the image with minimum error ,like this i want to do for 5 images

please help

Answer 1

[Sean de Wolski]

Use the second output argument from min() and the correct dim input to min() to tell you where the minimum is.

[kash]

Sean i cant understand ,please can you explain it please ,if possible with an example

Answer 2

[Sargondjani]

i have difficulty understanding your code... but lets say you have: w(:,:,1:5)

you can add another loop (iw=1:5;), store the errors in a vector: Er(iw,1).

Then

[value,index]=min(Er); %if Er is not a vector, then add the dimension (like Sean says)

w(:,:,index) gives you the matrix that minimized the error

something like that

[kash]

Sargondjani ,It is a for loop for wher e5 images are processed one by one,where each image is multiplied the matrix A1_20(:,:,j1),for example let us take first image,it is multiplied with A1_20(:,:,j1),so i get 20 images ,now in this i want to find the image which has minimum error and dispay,like this i have to do for 5 images,i cant understand why u used w(:,:,1:5),please provide assuistance

[Sargondjani]

i didnt understand much of your question, really. because you said you 20 images, but you want to do it for 5 images.... to confusing for me...

so let's say we change 5 into 20, do you get it then?

i mean, you have to store the 20 images + errors. say im=1:20, so you have images(:,:,im) and errors(im,1).

then pick the minimum error with:

[value,index]=min(error)

now you have the index, so you just have to retrive the corresponding image: image(:,:,index)

do you understand the method???

[kash]

Sargondjani i think i did not explain you well,

1.i have 5 images,i am passing it to a loop,one by one

2.ten i am multiplying each image by the matrix A1_20(:,:,j1),where this matrix contains 20 matrix values of zeros and ones,

3.so for 1st image i get 20 matices.(images),in this i want to find the image having minimum error ..

4.finally i want to do for all 5 images

[Sargondjani]

do you understand my method??

i mean, i prefer not to spell out solutions, but to give exampls so you can figure out the details yourself :-)