Tensor completion is one of the ecient methods for restoration of data. Tensor completion problem by minimization of the rank of tensor, leads to an appropriate solution, however, it gives a non-convex objective function that generates an NP-hard problem. In order to be solved this problem, at rst, by matricization it is transformed to some matrix completion problems. Next, instead of using the matrix rank, the trace norm is applied. In this paper, to increase the speed and accuracy of simple low rank tensor completion algorithm, a fast tri-factorization method is presented.
