RIP Conditions Based on Non-convex Optimization Model for Tensor Completion
The restricted isometric condition is of great importance in sparse optimization as a guarantee of sparsity.Many results have been obtained in compressed perception and matrix completion based on l1,L*,l1-l2 and L*-LF optimization models with restricted isometric conditions.In this paper,we extend them to tensors,and study the restricted isometry property(RIP)recovered from low-rank tensors X and give a bound on the constant δ2rn,based on the Tucker rank and L*-LF optimization models.
万方数据
维普
