Bi Sparsity Pursuit: A Paradigm for Robust Subspace Recovery
Technical Report,27 Sep 2015,30 Sep 2015
North Carolina State University Raleigh United States
Pagination or Media Count:
The success of sparse models in computer vision and machine learning is due to the fact that, high dimensional datais distributed in a union of low dimensional subspaces in many real-world applications. The underlying structuremay, however, be adversely affected by sparse errors. In this paper, we propose a bi-sparse model as a framework toanalyze this problem, and provide a novel algorithm to recover the union of subspaces in presence of sparsecorruptions. We further show the effectiveness of our method in a number of applications using real-world visiondata.