MRF-based rapid image restoration method

A restoration method and image technology, which is applied in the fields of cultural relics protection, computer image processing, and film and television special effects production, can solve the problems of large amount of calculation and low execution efficiency of algorithms, reduce running time, reduce computing time, and speed up message transmission and convergence speed effect

Active Publication Date: 2019-05-24
TIANJIN UNIV
View PDF4 Cites 0 Cited by
  • Summary
  • Abstract
  • Description
  • Claims
  • Application Information

AI Technical Summary

Problems solved by technology

This algorithm solves the problem of structure propagation better, but the algorithm has a large amount of calculation and low execution efficiency

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
  • MRF-based rapid image restoration method
  • MRF-based rapid image restoration method
  • MRF-based rapid image restoration method

Examples

Experimental program
Comparison scheme
Effect test

Embodiment 1

[0036] The embodiment of the present invention proposes a fast image restoration method based on MRF, such as figure 1 As shown, the specific steps and principles are as follows:

[0037] 101: Perform "pre-repair" on the low-resolution image at the top layer of the Gaussian pyramid, and perform upsampling and interpolation on the "pre-repair" result to obtain the preprocessed image I';

[0038] First, set the original image I as the high-resolution image G at the bottom of the pyramid 0 , which is down-sampled to obtain the top-level low-resolution image G K , where K represents the number of pyramid layers.

[0039] Second, using an adaptive sample block repair algorithm [2] for image G K Repair and get the pre-repair image G' K ; for the pre-restored image G' K Perform upsampling interpolation to obtain the initial repair image G' 0 (dimensions and G 0 same). Finally, the pixels in the damaged area in the original image I are replaced by the initial inpainted image ...

Embodiment 2

[0056] In order to obtain the initial information of the internal nodes of the MRF, this method proposes a "pre-repair" preprocessing method to achieve this goal. The schematic diagram of the preprocessing method is shown in Figure 4 (a) and Figure 4 As shown in (b): Figure (a) is a schematic diagram of the input damaged image I, in which φ is the source area, Ω is the damaged area, the wavy line in the dotted square represents known information, and the white area represents unknown information; b) is a schematic diagram of the preprocessed image I', and the black dots on the white background represent the initial information obtained after preprocessing. The present invention selects an image to display the result of the preprocessing method, such as Figure 4 (c) and Figure 4 As shown in (d): Figure (c) is the damaged image I instance "horse racing image", and the smeared part represents the damaged area. Figure (d) is the result of the preprocessed image I' instance. ...

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 a rapid image restoration method based on MRF. The method comprises the following steps: selecting a node with the highest priority from unaccessed nodes by utilizing an improved confidence coefficient propagation algorithm, marking the node as an accessed node, calculating messages of the node and all unaccessed adjacent nodes, and updating the confidence coefficient and the priority of the current adjacent node until all the nodes are marked as accessed nodes; and recovering the current node to be unaccessed according to the accessed reverse sequence, calculating messages of the node and all accessed adjacent nodes thereof, and updating the confidence coefficient and priority of the current adjacent node until all the nodes are recovered to be unaccessed. According to the method, the initial value of the MRF internal node in the damaged area is roughly estimated by using the pre-repair result of the low-resolution image, the candidate block of the MRF node isscreened by using the improved confidence coefficient calculation formula, and finally, the optimal matching block is determined by using the MRF, so that the repair effect and the operation time canbe effectively improved.

Description

technical field [0001] The invention relates to the field of computer image processing, in particular to a fast image repair method based on MRF (Markov Random Field), which can be widely used in the fields of cultural relic protection, film and television special effect production and the like. Background technique [0002] Digital image restoration technology is a research hotspot in the field of computer vision, and has been widely used in cultural relics protection, film and television production, old photo restoration, object removal, etc. [0003] Digital image inpainting techniques can be roughly divided into two categories: 1) Small area image inpainting techniques based on partial differential equations; 2) Texture synthesis techniques in larger damaged areas based on sample blocks. [0004] Image restoration based on partial differential equations mainly uses heat diffusion equations to establish partial differential equations of images. Traditional texture synthe...

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
Patent Type & Authority Applications(China)
IPC IPC(8): G06T5/00
Inventor 何凯沈成南刘坤刘志国
Owner TIANJIN UNIV
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
Try Eureka
PatSnap group products