Looking for breakthrough ideas for innovation challenges? Try Patsnap Eureka!

Blind Image Restoration Method Based on Blur Kernel Estimation Iterative Structure Preservation

An iterative structure and blur kernel technology, applied in the field of image processing, can solve the problems of obvious ringing, slow convergence, unclear texture detail recovery, etc., to achieve the effect of SSIM improvement

Active Publication Date: 2022-04-26
NORTHWESTERN POLYTECHNICAL UNIV
View PDF7 Cites 0 Cited by
  • Summary
  • Abstract
  • Description
  • Claims
  • Application Information

AI Technical Summary

Problems solved by technology

[0004] The technical problem solved by the present invention is: the existing single image blind restoration method has significant ringing in the blur kernel estimation process, which cannot be distinguished from the effective edge structure, resulting in unclear restoration of texture details, poor robustness and slow convergence. question

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
  • Blind Image Restoration Method Based on Blur Kernel Estimation Iterative Structure Preservation
  • Blind Image Restoration Method Based on Blur Kernel Estimation Iterative Structure Preservation
  • Blind Image Restoration Method Based on Blur Kernel Estimation Iterative Structure Preservation

Examples

Experimental program
Comparison scheme
Effect test

Embodiment Construction

[0056] The technical method of the present invention includes the following parts: establishing an image degradation model, estimating a potentially sharp image, selecting a structure-preserving gradient subset, and estimating a blur kernel. details as follows:

[0057] 1) Establish an image degradation model

[0058] A degradation model is established for the input blurred image, and the potential image gradient and blur kernel are constrained at the same time, and the blur kernel and the potential clear image are simultaneously solved by alternate iterations.

[0059] 2) Estimation of potentially sharp images

[0060] According to the degradation model of the image, at each step of the iterative solution, fix the blur kernel and add L to the latent image gradient 1 Regularization constraints are introduced, and variables are introduced to solve the latent image using iterative re-weighted least squares (IRLS), and obtain the intermediate latent image result sequence when t...

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 blind image restoration method based on fuzzy kernel estimation iterative structure preservation. The method firstly extracts the change characteristics of the latent image sequence in the kernel estimation iteration, and performs difference calculation on the latent image results of two adjacent iterations to obtain the image Sequence difference result map; Threshold selection is performed according to the difference result, and the favorable structure in each image block is screened out by dividing the difference image into blocks; For the gradient map of the current potential image, the favorable structure of the image is extracted according to the screening result Gradients to obtain a subset of image gradients; then based on the extracted image gradients, estimate the blur kernel to obtain more accurate estimation results. The present invention can well restore clear images with more details. Experiments show that compared with the original blurred image, PSNR and SSIM of the image restored by the present invention are increased by 15% and 37.3%.

Description

technical field [0001] The invention belongs to the field of image processing, and relates to an image blind restoration method, in particular to an image blind restoration method based on fuzzy kernel estimation iterative structure preservation. Background technique [0002] The quality of image blind restoration depends on the quality of the estimated blur kernel. In order to estimate the blur kernel more accurately, it is necessary to find useful structural information for kernel estimation in the image. Existing image restoration methods are mainly based on image priors and methods based on image salient edge extraction to find favorable structures in images. [0003] Document "Hu Fuyuan, Wang Zhenhua, Lv Fan, etc. A motion blur image restoration method based on saliency edges [J]. Journal of Suzhou Institute of Science and Technology (Natural Science Edition), 2017, 34(1): 77-82." A saliency edge-based motion blur restoration method is proposed, which combines image gr...

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 Patents(China)
IPC IPC(8): G06T5/00
CPCG06T2207/20056G06T5/77
Inventor 朱宇张艳宁白雪孙瑾秋
Owner NORTHWESTERN POLYTECHNICAL UNIV
Who we serve
  • R&D Engineer
  • R&D Manager
  • IP Professional
Why Patsnap Eureka
  • Industry Leading Data Capabilities
  • Powerful AI technology
  • Patent DNA Extraction
Social media
Patsnap Eureka Blog
Learn More
PatSnap group products

Browse by: Latest US Patents, China's latest patents, Technical Efficacy Thesaurus, Application Domain, Technology Topic, Popular Technical Reports.

© 2024 PatSnap. All rights reserved.Legal|Privacy policy|Modern Slavery Act Transparency Statement|Sitemap|About US| Contact US: help@patsnap.com