Multi-filters guided low-rank tensor coding for image inpainting
作者：
Highlights：
• The method introduces the multi-filters guided low-rank tensor coding as a prior information for image inpainting.
• We first convolve the target image with FOE filters to formulate multi-feature images, and then regard the extracted similarity grouped cube as a low-rank tensor.
• The resulting non-convex model is iteratively tackled by gradient descent procedure.
• An aggregation version of proposed method is explored for further improving the inpainting performance.
摘要
•The method introduces the multi-filters guided low-rank tensor coding as a prior information for image inpainting.•We first convolve the target image with FOE filters to formulate multi-feature images, and then regard the extracted similarity grouped cube as a low-rank tensor.•The resulting non-convex model is iteratively tackled by gradient descent procedure.•An aggregation version of proposed method is explored for further improving the inpainting performance.
论文关键词：Image inpainting,Multi-filters,Low-rank,Tensor coding,Higher-order singular value decomposition,Gradient descent,Weighted aggregation
论文评审过程：Received 12 February 2018, Revised 11 September 2018, Accepted 17 September 2018, Available online 4 October 2018, Version of Record 12 March 2019.
论文官网地址：https://doi.org/10.1016/j.image.2018.09.010
