This work deals with recovering a low rank matrix from its lower dimensional projections via nuclear norm minimization.
% Minimize ||X||* (nuclear norm of Z)
% Subject to A(X) = Y
We use split Bregman algorithm for the same.
% Minimize (lambda1)||W||* + 1/2 || A(X) - y ||_2^2 + eta/2 || W-X-B1 ||_2^2
%W is proxy variable and B1 is the Bregman variable
The use of Bregman technique improves the convergence speeds of our algorithm and gives a higher success rate. Also, the accuracy of reconstruction is much better even for cases where small number of linear measurements are available.
Download apps, toolboxes, and other File Exchange content using Add-On Explorer in MATLAB.