(L2,L1)-TRIP Conditions Based on||x||*-α||x||F Model Tensor Recovery
The limiting isometric condition is a sparsity guarantee in sparse optimisation.In compressed sensing and matrix complementation,the restriction isometry condition based on L1,L*,L1-L2 and L*-LF optimisation models has been extended to tensor models.In this paper,we study the generalised(L2,Lt)-restriction isometry property((L2,L1)-TRIP)of low rank tensor X recovery based on the t-rank and L*-αLF(0＜α≤1)optimisation models in the presence of impulse noise,and give sufficient conditions for the recovery of the low rank tensor x and the recovered error.
万方数据
