Dual Principal Component Pursuit for Robust Subspace Learning:
          Theory and Algorithms for a Holistic Approach

               Tianyu Ding 1 Zhihui Zhu 2 Rene Vidal 3 Daniel P. Robinson 4

             Abstract                 outliers and noise. Unlike inliers, which exactly lie in the
   The Dual Principal Component Pursuit (DPCP)          subspace, outliers are far from the subspace and do not ex-
   method has been proposed to robustly recover         hibit linear structure. Noise i ...