A nonconvex formulation for low rank subspace clustering: algorithms and convergence analysis (Q1639712)

The authors obtain a low rank subspace clustering method based on non-convex modeling in order to give a solution to the problem of subspace clustering with data that is potentially corrupted by both dense noise and sparse gross errors. Refined smoothness properties of the spectral function are derived and used to establish the convergence of the alternating direction method of multipliers proposed for solving the non-convex problem. It is proved that the iterates are uniformly bounded and all limit points verify the first-order optimality conditions. Numerical experiments prove the efficiency of the technique. They process real data from face and digit clustering and motion segmentation.