Convolution in MATLAB

                 Let us try to understand convolution by performing spatial averaging on a matrix without using MATLAB built in function ‘conv2()’.

%CONVOLUTION IN MATLAB

%INPUT  MATRIX
A = zeros(5);
A(:) = 1:25;

%KERNEL
avg3 = ones(3)/9;

%CONVOLUTION
Result = conv2(A,avg3,'same');
display(Result);

Steps to be performed:





NOTE :
To define a kernel for spatial averaging, fill the kernel with ones and divide it by the number of elements in it.
For instance, consider kernel of size 4x4 , fill the  matrix with ones and divide it by 16. i.e the total number of elements in the matrix.

MATLAB CODE:
%INPUT  MATRIX
A = zeros(5);
A(:) = 1:25;

%KERNEL
avg3 = ones(3)/9;

%PAD THE MATRIX WITH ZEROS
B = padarray(A,[1 1]);
% PRE-ALLOCATE THE MATRIX
Output = zeros([size(A,1) size(A,2)]);

%PERFORM COONVOLUTION
for i = 1:size(B,1)-2
    for j = 1:size(B,2)-2
        Temp = B(i:i+2,j:j+2).*avg3;
        Output(i,j) = sum(Temp(:));
    end
 end

display(Output);

10 comments:

  1. UnknownNovember 4, 2015 at 3:37 PM

    Its interesting and good for students.

    ReplyDelete
  2. Jacob martinNovember 10, 2015 at 12:17 PM

    The key is to give the peruser a little look into your experience, which urges them to need to take in more by perusing your resume
    Medical Science Homework Help

    ReplyDelete
  3. VictorJune 4, 2016 at 1:37 AM

    Congratulations very interesting and didactically, do you have a youtube channel? if no, think about it.

    ReplyDelete
  4. AnonymousNovember 4, 2016 at 4:11 AM

    Man You are a life save, thanks for this easy explaining :) :)

    ReplyDelete
  5. AnonymousNovember 7, 2016 at 6:21 AM

    Hey In understand your example very clearly. I have question. If i want to subtraction any pixel intensity from its 8 neighbors pixel intensity(this is for every pixels in the image) and then take the absolute sum of the subtraction. How should I define the kernel (avg3 in your example) and perform the task with conv2?

    ReplyDelete
  6. UnknownNovember 10, 2016 at 4:09 AM

    Thanks was very helpful for my assignment!

    ReplyDelete
  7. UnknownApril 14, 2017 at 9:17 PM

    plz tell me how to use it , I cant understand...

    ReplyDelete
  8. fray_dogApril 23, 2018 at 2:04 AM

    Hello, I keep getting the following error "Subscript indices must either be real positive integers or logicals.

    Error in test (line 16)
    Output(i,j) = sum(Temp(:));"

    do you have any advice? thank you

    ReplyDelete
  9. MicheleJune 6, 2018 at 1:25 PM

    I tried with a kernel containing elements with different values (not a constant value as in your example) and the convolution with conv2 or with the for loop gave different results, why? I used the following code:

    %CONVOLUTION IN MATLAB with conv2
    clear
    %INPUT MATRIX
    A = zeros(5);
    A(:) = 1:25;

    %KERNEL
    avg3 = rand(3);

    %CONVOLUTION
    Result = conv2(A,avg3,'same');
    display(Result);



    %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
    % EXPLICIT calculation using the same kernel


    %PAD THE MATRIX WITH ZEROS
    B = padarray(A,[1 1]);
    % PRE-ALLOCATE THE MATRIX
    Output = zeros([size(A,1) size(A,2)]);

    %PERFORM COONVOLUTION
    for i = 1:size(B,1)-2
    for j = 1:size(B,2)-2
    Temp = B(i:i+2,j:j+2).*avg3;
    Output(i,j) = sum(Temp(:));
    end
    end

    display(Output);

    ReplyDelete
  10. Aaron AngelJune 23, 2018 at 2:09 AM

    @Michele

    Rotate the convolution mask by 180 degrees.

    Example:
    Temp = B(i:i+2,j:j+2).*rot90(avg3,2);

    The mask in the given example is symmetric so rotating it by 180 degree yielded the same mask.

    ReplyDelete