Can any body tell me how to implement KSVD in matlab?

i am doing the project on "quaternion matrix analysis of color image using vector sparse model" so i need to implement ksvd so can please anynoe help me???

Answer 1

[Royi Avital]

Here is a function which implements the K-SVD algorithm.

function [ mD ] = DL_KSVD( mD, mX, paramCardinality )

      numberOfAtoms = size(mD, 2);

      for ii = 1:50
          %<! Update the Representations
          mA = Omp(mX, mD, paramCardinality);

          %<! Update the Dictionary
          for jj = 1:5
              mE = mX - (mD * mA);
              for kk = 1:numberOfAtoms
                  vP         = find(mA(kk, :));
                  mEP        = mE(:, vP) + mD(:, kk) * mA(kk, vP);
                  vA         = mD(:, kk)' * mEP;
                  mA(kk, vP) = vA;
                  mD(:, kk)  = mEP * vA' / (vA * vA');
              end
          end

          mD = bsxfun(@rdivide, mD, sqrt(sum(mD .^ 2, 1)));

          %%Progress (Debug):
  %         mA     = omp(mD' * mX, mD' * mD, cardinality);
  %         mX_hat = mD * mA;
  %         A = mean(abs(mX(:) - mX_hat(:)))

      end

end

The Omp() function mA = Omp(mX, mD, paramCardinality) solves mD * mX - mA_2 s.t. each column of mA has cardinality (L0 Pseudo Norm) less equal to paramCardinality. It uses the Orthogonal Matching Pursuit method.

[RS SHARMA]

can i have function omp(mX.mD,cardinality). cardinality means what value..give example.

Answer 2

[murat simsek]

What is the meaning of "mD" parameter?