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

Moving image deblurring method based on self-adaptive residual errors and recursive cross attention

A moving image, attention technology, applied in image enhancement, image analysis, image data processing and other directions, can solve the problem of motion blurred image non-uniformity and so on

Active Publication Date: 2021-01-01
GUILIN UNIV OF ELECTRONIC TECH
View PDF4 Cites 4 Cited by
  • Summary
  • Abstract
  • Description
  • Claims
  • Application Information

AI Technical Summary

Problems solved by technology

This method can solve the non-uniformity problem of motion blur images, remove artifacts and obtain more high-frequency features of images, and reconstruct high-quality images with rich texture details

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
  • Moving image deblurring method based on self-adaptive residual errors and recursive cross attention
  • Moving image deblurring method based on self-adaptive residual errors and recursive cross attention
  • Moving image deblurring method based on self-adaptive residual errors and recursive cross attention

Examples

Experimental program
Comparison scheme
Effect test

Embodiment

[0053] refer to figure 1 , a moving image deblurring method based on adaptive residual and recursive cross-attention, comprising the following steps:

[0054] 1) Establishment of the defuzzification network framework: based on the idea of ​​adversarial games, the defuzzification network includes a generation network G and a discrimination network D, and the generation network G is equipped with a shallow feature extraction module M e , adaptive residual module namely ARM, recursive cross-attention module namely RCCAM and feature reconstruction module M r , the discriminant network D discriminates the learned deblurred image and clear image;

[0055] 2) Shallow feature extraction: In step 1), the input of the generation network G is the original blurred image B,

[0056] First use the M e Perform shallow feature extraction on the input blurred image B to obtain shallow feature P 0 As shown in formula (1):

[0057] P 0 = M e (B)(1);

[0058] 3) Adaptive residual process:...

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 moving image deblurring method based on self-adaptive residual error and recursive cross attention, which is characterized by comprising the following steps of: 1) establishing a deblurring network framework; (2) carrying out shallow feature extraction, (3) carrying out an adaptive residual process, (4) carrying out a recursive cross attention process, (5) carrying out image reconstruction and (6) carrying out network model discrimination. The method can solve the non-uniformity problem of a motion blurred image, remove artifacts, obtain more image high-frequency features and reconstruct a high-quality image with rich texture details.

Description

technical field [0001] The invention relates to the technical field of intelligent image processing, in particular to a moving image deblurring method based on adaptive residual and recursive cross attention. Background technique [0002] Motion blurred images are degraded images caused by camera shake and object movement when taking pictures. The purpose of moving image deblurring is to reconstruct and estimate an unknown clear image given a motion blurred image. This application is beneficial to other computer vision tasks, such as object retrieval, image restoration, action recognition, event detection, image quality evaluation, etc. [0003] In recent years, thanks to the superior performance of deep learning in image restoration, it has been used in motion image deblurring applications to achieve good results. Gong et al. expressed non-uniform motion blur as pixel-wise linear motion blur, and the proposed method utilizes a fully-convolutional deep Neural Network (Full...

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): G06T5/00G06N3/04G06N3/08
CPCG06N3/08G06T2207/10016G06T2207/20081G06T2207/20084G06N3/048G06N3/045G06T5/73
Inventor 欧阳宁邓超阳李祖锋林乐平莫建文袁华
Owner GUILIN UNIV OF ELECTRONIC TECH
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