Eureka AIR delivers breakthrough ideas for toughest innovation challenges, trusted by R&D personnel around the world.

Efficient image restoration method based on alternating direction method

A technology of alternating direction method and repair method, which is applied in the field of high-efficiency image repair based on alternating direction method, can solve the problems of low repair efficiency and slow convergence speed, and achieve the effect of improving repair accuracy, improving completion speed, and improving parallelism.

Inactive Publication Date: 2014-10-29
ZHEJIANG UNIVERSITY OF MEDIA AND COMMUNICATIONS
View PDF3 Cites 11 Cited by
  • Summary
  • Abstract
  • Description
  • Claims
  • Application Information

AI Technical Summary

Problems solved by technology

Compared with the repair method based on local information filling, the calculation result of LRTC method is accurate, the repair precision is high, and the effect is good, but its convergence speed is very slow, resulting in low repair efficiency
[0005] To sum up, the existing image inpainting methods have a contradiction between inpainting quality and inpainting efficiency, and new methods are needed to achieve efficient and perfect image inpainting

Method used

the structure of the environmentally friendly knitted fabric provided by the present invention; figure 2 Flow chart of the yarn wrapping machine for environmentally friendly knitted fabrics and storage devices; image 3 Is the parameter map of the yarn covering machine
View more

Image

Smart Image Click on the blue labels to locate them in the text.
Viewing Examples
Smart Image
  • Efficient image restoration method based on alternating direction method
  • Efficient image restoration method based on alternating direction method
  • Efficient image restoration method based on alternating direction method

Examples

Experimental program
Comparison scheme
Effect test

Embodiment Construction

[0045] The present invention will be further described in detail below in conjunction with specific embodiments.

[0046] An efficient image inpainting method based on the alternating direction method, including:

[0047] (1) Decompose the original image into three low-rank matrices according to the R / G / B channel;

[0048] (2) Divide each low-rank matrix into several block matrices, and use the alternating direction method to complete each block matrix to obtain a completed block matrix.

[0049] In this embodiment, the KD tree method is used to divide each low-rank matrix into several block matrices, which specifically include the following steps:

[0050] (2-11) Construct the KD tree of the current low-rank matrix;

[0051] (2-12) Use the part of the current low-rank matrix corresponding to each node in the KD tree as each block matrix of the current low-rank matrix.

[0052] The number of block matrices corresponding to each low-rank matrix and the size of each block mat...

the structure of the environmentally friendly knitted fabric provided by the present invention; figure 2 Flow chart of the yarn wrapping machine for environmentally friendly knitted fabrics and storage devices; image 3 Is the parameter map of the yarn covering machine
Login to View More

PUM

No PUM Login to View More

Abstract

The invention discloses an efficient image restoration method based on an alternating direction method. The efficient image restoration method based on the alternating direction method includes: translating image restoration problems into low rank matrix complement problems; further dividing each low rank matrix into a plurality of partitioned matrixes by decomposing an original image into three low rank matrixes according to R / G / B channels; complementing each partitioned matrix; combining all the partitioned matrixes after being complemented so as to obtain corresponding complement matrixes; further obtaining a repaired image according an combination of the complement matrixes. The efficient image restoration method based on the alternating direction method facilitates parallel processing through partitioning, complements each partitioned matrix by using the alternating direction method, greatly improves image restoration speed, can not damage overall relevance of the original matrixes due to the fact that the plurality of the partitioned matrixes are obtained by dividing the three low rank matrixes according to relative independence of the three low rank matrixes, and facilitates improvement of restoration accuracy. The efficient image restoration method based on the alternating direction method not only can be used for image restoration, but also can be used in fields of object removal, image noise reduction and the like.

Description

technical field [0001] The invention relates to the field of image processing, in particular to an efficient image restoration method based on an alternating direction method. Background technique [0002] Image inpainting is to restore the complete image through various methods based on the known parts of the image. At present, there are many methods, which can be divided into two categories: filling based on local information and approximation completion based on global information. [0003] Methods based on PDE algorithm (Partial differential equation, partial differential equation) and BP algorithm (Back Propagation, backpropagation) focus on local relations. The PDE algorithm tries to fill the lost area by deriving from the boundary of the lost area to the inside. The latter category first establishes a Markov random field (MRF) on the input image, and then performs global optimization through information exchange between adjacent nodes. Both types of algorithms assu...

Claims

the structure of the environmentally friendly knitted fabric provided by the present invention; figure 2 Flow chart of the yarn wrapping machine for environmentally friendly knitted fabrics and storage devices; image 3 Is the parameter map of the yarn covering machine
Login to View More

Application Information

Patent Timeline
no application Login to View More
IPC IPC(8): G06T5/00
Inventor 张根源
Owner ZHEJIANG UNIVERSITY OF MEDIA AND COMMUNICATIONS
Who we serve
  • R&D Engineer
  • R&D Manager
  • IP Professional
Why Eureka
  • Industry Leading Data Capabilities
  • Powerful AI technology
  • Patent DNA Extraction
Social media
Eureka Blog
Learn More
PatSnap group products