Image Deblurring Method Based on Compression and Excitation Mechanism Neural Network

A technology of neural network and incentive mechanism, applied in biological neural network model, neural learning method, image enhancement, etc., to achieve high portability, optimized processing process, and excellent effect of clear image quality

Active Publication Date: 2022-06-21
SUN YAT SEN UNIV
View PDF4 Cites 0 Cited by
  • Summary
  • Abstract
  • Description
  • Claims
  • Application Information

AI Technical Summary

Problems solved by technology

[0003] It has been a while since the invention of the replacement module and structure used in the above technology. At present, more researchers have proposed more novel and efficient network modules, so the above technology still has a lot to do in restoring the quality of clear images. big room for improvement

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
  • Image Deblurring Method Based on Compression and Excitation Mechanism Neural Network
  • Image Deblurring Method Based on Compression and Excitation Mechanism Neural Network
  • Image Deblurring Method Based on Compression and Excitation Mechanism Neural Network

Examples

Experimental program
Comparison scheme
Effect test

Embodiment

[0058] like figure 1 As shown, the image deblurring method based on the compression and excitation mechanism neural network of the present invention includes the following steps:

[0059] S1. Obtain the data set required for training the network; the data set is a plurality of image pairs, each image pair is composed of a blurred image and its corresponding clear image; the data set is obtained according to the specific scene of actual use, for example, when it is necessary to process a high-speed camera When capturing a blurred license plate image, it is necessary to construct a dataset of high-speed moving cars and license plates according to this situation; the dataset should be as large as possible to ensure that the dataset can cover the corresponding blurring patterns.

[0060] In this embodiment, when constructing the data set, the minimum resolution of the blurred image is 256*256, because a 256*256 block in the image pair will be randomly cropped for training during n...

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 deblurring method based on a compression and excitation mechanism neural network, comprising the following steps: obtaining a data set required for training the network, the data set being a plurality of image pairs, each image pair consisting of a blurred image and its Corresponding clear image composition; construct compression and excitation defuzzification network, the network is a multi-scale network, each scale has the same structure, including encoder-decoder structure and ConvLSTM layer; use data set to compress and excitation defuzzification network Perform training; use the trained compression and excitation deblurring network to process the blurred image. The present invention improves the residual block in the feature processing module on the basis of the SRN defuzzification network, introduces a compression and incentive mechanism, and obtains the SE residual block applied to the network of the present invention, and then constitutes the SEDN defuzzification network, so that the final The restored clear image quality is even better.

Description

technical field [0001] The invention belongs to the technical field of computer vision and image processing, and relates to an image deblurring method based on a compression and excitation mechanism neural network. Background technique [0002] Using multi-scale convolutional neural networks to deal with image deblurring has been a problem that has been continuously studied and improved by many researchers in recent years. In 2018, Tao et al. improved the SRN deblurring network mentioned in their conference paper "Scale-recurrentnetwork for deep image deblurring". The SRN deblurring network is based on an earlier multi-scale convolutional neural network for image deblurring. By replacing the original simple network modules and structures with some higher-performance network modules and structures in the original simpler network, the SRN deblurring network has achieved a significant improvement in the quality of the clear images recovered by the network. [0003] It has bee...

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): G06T5/00G06N3/04G06N3/08
CPCG06T5/003G06N3/084G06N3/045
Inventor 李洽吴佳琪
Owner SUN YAT SEN 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

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

© 2024 PatSnap. All rights reserved.Legal|Privacy policy|Modern Slavery Act Transparency Statement|Sitemap