Abstract: Block-coded videos suffer from the blocking artifacts after being decoded.In order to solve this problem,a block-based deblocking method using low-rank tensor recovery was proposed.First,three order tensor is constructed through clustering similar blocks in video sequence.Then,according to low-rank property of background tensor and sparsity of blocking artifacts,the proposed approach utilizes the augmented Lagrange multiplier method which extends to tensor to solve the low-rank tensor recovery problem.The proposed approach utilizes tensor model to preserve the structural properties of high dimensional data.Experimental results show that it can obtain higher PSNR value and better visual effect comparing with traditional deblocking methods.

Key words: Video codec,Video deblocking,Tensor recovery,Augmented Lagrange multiplier