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

Blind image deblurring method based on recurrent multi-scale generative adversarial network

A motion blur and multi-scale technology, applied in the field of image processing, can solve the problems affecting the restoration effect and the accuracy of blur kernel estimation, and achieve the effect of easy training, omitting the blur kernel estimation process, and good restoration effect

Inactive Publication Date: 2021-04-09
HANGZHOU DIANZI UNIV
View PDF2 Cites 0 Cited by
  • Summary
  • Abstract
  • Description
  • Claims
  • Application Information

AI Technical Summary

Problems solved by technology

In the blind deblurring of a single moving image, the blur kernel and its size of the blurred image are unknown, which will affect the accuracy of the blur kernel estimation, and then affect the final restoration effect

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 deblurring method based on recurrent multi-scale generative adversarial network
  • Blind image deblurring method based on recurrent multi-scale generative adversarial network
  • Blind image deblurring method based on recurrent multi-scale generative adversarial network

Examples

Experimental program
Comparison scheme
Effect test

Embodiment Construction

[0037] The specific implementation of the present invention will be further described below.

[0038] The fuzzy image set B is input into the generator G, and the generator output image set L is obtained, which is used as the input of the discriminator D, and the discrimination result of the discriminator is obtained. Similarly, the clear image set S is also used as the input of the discriminator to obtain the discriminant result. The judgment result indicates whether the input is from the clear image set or the generated image set. If the judgment result is greater than 0.5, it is judged as the clear image set S; otherwise, it is judged as the generator output image set L. Calculate the error between the judgment result and the real label data, use the gradient descent algorithm to optimize the discriminator, then calculate the error mean of the generated image and the clear image, and use the gradient descent algorithm to optimize the generator. Alternately optimize the dis...

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 image blind motion blur removal method based on a cyclic multi-scale generation confrontation network. The method of the invention uses a circular multi-scale encoder and a decoder as a generator, and constructs a corresponding decision device. The adversarial loss of generated images and clear images, multi-scale mean square error and multi-scale gradient error are used as the loss function of the generative adversarial network, and the loss function is optimized by gradient descent method. The invention uses the generation confrontation network to learn the relationship between the motion blur image and its corresponding clear image, which saves the complicated fuzzy kernel estimation process. The method of the invention can extract the edge features of the image, has a simpler network structure and fewer parameters, and the network model is easier to train and has a better restoration effect.

Description

technical field [0001] The invention belongs to the technical field of image processing, and relates to an image blind motion blurring method based on a cyclic multi-scale generation confrontation network. Background technique [0002] Since it is difficult to maintain a relatively still state between the photographing device and the imaged object, motion blur in the image is caused. However, in daily life, traffic safety, medicine, military investigation and other fields, it is particularly important to be able to obtain a clear image. [0003] The blurring of moving images can be regarded as the convolution of a clear image and a two-dimensional linear function, which is polluted by additive noise. This linear function is called the point spread function or convolution kernel, and it contains the blurring information of the image. Blind deblurring of images refers to restoring the original clear image only by relying on the information of the blurred image itself when th...

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): G06T5/00G06T7/13G06K9/62G06N3/04G06N20/00
CPCG06N20/00G06T7/13G06N3/045G06F18/214G06T5/73
Inventor 陈华华陈富成叶学义
Owner HANGZHOU DIANZI 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