Non-local image inpainting method based on low-rank matrix recovery

Abstract

The invention discloses a non-local image inpainting method based on low-rank matrix recovery. The method comprises the following steps: pre-completing a low-rank texture and a natural image respectively; partitioning the pre-completed image into a plurality image blocks, and calculating the matching block matrix of each image block on the basis of block matching and a grouping method; inpainting each image block in the matching block matrix by using low-rank matrix completion; integrating the repairing results of all the image blocks, and repairing an image to be repaired according to an integration result. Two image completion methods based on examples and numerical calculation are used comprehensively. The method has a wider application scene range than an existing image completion method. In particularly, the image can be reconstructed from random sampling in an almost perfect way. Moreover, pre-completion is performed firstly, and block completion is performed secondly in a grouping way, so that the repairing accuracy is increased greatly. The method can be conveniently applied to a plurality of practical applications such as rendering acceleration, image inpainting and object removal.

Description

technical field [0001] The invention relates to the field of computer images, in particular to a non-local image restoration method based on low-rank matrix restoration. Background technique [0002] Image inpainting algorithm has always been one of the difficult and hot issues in the image field. At present, the methods with better results are divided into two categories: methods based on numerical calculations and methods based on examples. [0003] Numerical calculation-based methods include PDE methods and TV methods. PDE-based methods try to imitate the behavior of experts in the image restoration process: fill in the missing parts in the image by spreading Laplacians in the isoray direction. Total Variation (TV), first proposed as an effective denoising tool, this method was soon extended to handle more applications including image inpainting, deblurring and super-resolution. These two methods work in a similar way, and both are good solutions for smoothing the gradi...

AI Technical Summary

Problems solved by technology

[0005] In view of the shortcomings of the existing repair methods such as narrow application range, such as numerical calculation-based methods such as PDE and TV often cannot solve the problem of texture completion, the low-rank matrix restoration method is only effective for low-rank data, and the sample-based The method cannot repair pictures containing more random impulse noise. The present invention is a non-local image repair method based on low-rank matrix restoration. The non-local image repair method combines a sample-based method and a numerical calculation-based method. , combining the advantages of the two methods, it can repair images in various application scenarios (including scenes with random missing pixels) under the condition of ensuring the operation speed and accuracy

Images