A patch-based low-rank tensor approximation model for multiframe image denoising

作者：

Highlights：

• An algorithm for low-rank tensor approximation is proposed.

• The algorithm is based on matrix factorization to all-mode unfoldings of the tensor.

• The algorithm is embedded in a patch-based multiframe image denoising method.

• The performance of the denoising method is competitive in the numerical experiments.

摘要

•An algorithm for low-rank tensor approximation is proposed.•The algorithm is based on matrix factorization to all-mode unfoldings of the tensor.•The algorithm is embedded in a patch-based multiframe image denoising method.•The performance of the denoising method is competitive in the numerical experiments.

论文关键词：Low-rank tensor,Augmented Lagrangian alternating,Patch-based model,Image denoising

论文评审过程：Received 1 September 2016, Revised 17 November 2016, Available online 21 January 2017, Version of Record 17 October 2017.

论文官网地址：https://doi.org/10.1016/j.cam.2017.01.022