Robust subspace recovery via dual sparsity pursuit

A computer-implemented method of performing intensity variation segmentation using a dual sparse model framework, the method comprising:

• receiving a myocardial perfusion image sequence comprising a plurality of images depicting fluid passing through a cardiac structure over time;
• creating an image matrix based on a continuous image sequence;
• initializing a background matrix, a foreground matrix, and a coefficient matrix;
• performing a subspace recovery process over a plurality of iterations until convergence of one or more of the background matrix, the foreground matrix, and the coefficient matrix over the plurality of iterations, the subspace recovery process comprising:
• updating the background matrix based on the image matrix and the foreground matrix,
• updating the coefficient matrix by minimizing an L−1 norm of the coefficient matrix using a first linearized soft-thresholding process, and
• updating the foreground matrix by minimizing an L−1 norm of the foreground matrix using a second linearized soft-thresholding process;
• and
• using the foreground matrix to generate a measurement of intensity variation across the myocardial perfusion image sequence.

An article of manufacture for detecting a foreground data in an image sequence using a dual sparse model framework, the article of manufacture comprising a non-transitory, tangible computer-readable medium holding computer-executable instructions for performing a method comprising:

• creating an image matrix based on a continuous image sequence;
• initializing a background matrix, a foreground matrix, and a coefficient matrix;
• performing a subspace recovery process over a plurality of iterations until convergence of one or more of the background matrix, the foreground matrix, and the coefficient matrix over the plurality of iterations, the subspace recovery process comprising:
• update the background matrix based on the image matrix and the foreground matrix,
• update the coefficient matrix by minimizing an L−1 norm of the coefficient matrix using a first linearized soft-thresholding process, and
• update the foreground matrix by minimizing an L−1 norm of the foreground matrix using a second linearized soft-thresholding process;
• generating one or more background images based on the background matrix;
• and
• generating one or more foreground images based on the foreground matrix.