Dear All.
I'm Roy Nugroho from Indonesia.
Now I do my final project about Super Resolution Reconstruction.
In paper that i learn, there is a variable named D --> Decimation Matrix or
Decimation Operator, or Downsample Matrix.
Y(t) = DH(t)X(t) + N(t)
X(t) = High Resolution Image
size = (L^2 x 1) --> notation is in vector,
size of image X is (L x L), and reshape to (L^2 x 1)
Y(t) = Low Resolution Image
size = (M^2 x 1)--> notation is in vector,
size of image Y is (M x M), and reshape to (M^2 x 1)
H(t) = Blur Matrix
size = (L^2 x L^2)
This matrix is got by convmtx2 in Matlab. This matrix is Toeplitz.
N(t) = Gaussian Noise
size = (M^2 x 1)
D = Decimation Matrix
size = (M^2 x L^2)
As far as i know, D in Matlab is 'downsample'.
But 'downsample' in Matlab is different from above equation.
Would all of you who have the information about this matrix can explain to
me?
Thank you very much
Best Regard
Roy Nugroho
Institute Technology of Sepuluh Nopember
Surabaya - Indonesia