Applying Regression Model results on multispectral image

I used the regression learner application to develop a regression model. Then exported the model to the workspace.

I want to apply the exported model on a multispectral image.

I used foor loop to convert the image to table and apply the model. it works but it took very long time. I believe there is much faster way.

My code:

B = My_image;
a = size (B,1);
b = size (B,2);
c = size (B,3);
D = zeros(a,b);
l=0;
for i = 1:a
    for j = 1:b
        S.B01 = B(i,j,1);
        S.B05 = B(i,j,5);
        S.B06 = B(i,j,6);
        S.B07 = B(i,j,7);
        S.B09 = B(i,j,9);
        S.B10 = B(i,j,10);
        S.B11 = B(i,j,11);
        S.B12 = B(i,j,12);
        S.B02 = B(i,j,2);
        S.B03 = B(i,j,3);
        S.B04 = B(i,j,4);
        S.B08 = B(i,j,8);
        S.B8a = B(i,j,13);
        TT = struct2table( S );
        l = FDT.predictFcn(TT);
        D (i,j) = l;
    end
end

May anyone help me in that

Thank You

Answer 1

[Neha]

Hi Nagwa,

I understand that you want to convert a multispectral image to a table, so that it can be used in the regression model. Instead of using a nested loop to convert each pixel of the image to a table and apply the model, you can vectorize the process as vectorized code often runs much faster than the corresponding code containing loops.

For more information on vectorization, you can refer to the following documentation link:

https://www.mathworks.com/help/matlab/matlab_prog/vectorization.html

Hope this helps!