Low-illumination image classification method based on attention mechanism and capsule network

A low-light, attention-focused technology, applied in the field of image enhancement, image classification, and deep learning, can solve problems such as difficult to distinguish, lack of low-light processing, restoration and other image processing difficulties

Active Publication Date: 2020-11-17
GUILIN UNIV OF ELECTRONIC TECH
View PDF7 Cites 15 Cited by
  • Summary
  • Abstract
  • Description
  • Claims
  • Application Information

AI Technical Summary

Problems solved by technology

[0002] Due to the fact that in real life, there are a large number of images taken in low-light environments, a lot of information cannot be accurately obtained during use, and it is difficult to distinguish with the naked eye, which leads to difficulties in image processing such as image classification, target detection, and restoration.
While one major breakthrough after another has been achieved in image classification, most of them are dealing with bright images, and low light processing is noticeably lacking.

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
  • Low-illumination image classification method based on attention mechanism and capsule network
  • Low-illumination image classification method based on attention mechanism and capsule network
  • Low-illumination image classification method based on attention mechanism and capsule network

Examples

Experimental program
Comparison scheme
Effect test

Embodiment Construction

[0081] The specific embodiments of the present invention will be further described below in conjunction with the accompanying drawings, but the present invention is not limited.

[0082] figure 1An attention mechanism CBAM is shown, which is composed of a Channel attention module and a Spatial attention module. In the channel attention, each channel of the feature map is simultaneously globally pooled (via MaxPool) and average pooling (through AvgPool), each pixel of the feature map in the spatial attention is simultaneously subjected to global pooling (through MaxPool) and average pooling (through AvgPool), after these two attention The Sigmoid activation function of the force module is weighted to obtain a weighted feature map.

[0083] The attention mechanism was first introduced from biology in 1998, called salience, and then defined as attention in 2014. It was also used in machine translation earlier, and now it has become a neural network field. important concept. In...

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 low-illuminance image classification method based on an attention mechanism and a capsule network, and solves the technical problems that a low-illuminance image shot under weak light is too dark, so that the visual effect and the image classification result effect are poor, and detail loss and noise influence the classification of the image. According to the method, a CBAM attention mechanism network is utilized to enhance the brightness of a low-illumination image, remove the noise of the image and improve the details of the low-illumination image, and then the enhanced image is input into a capsule network for classification. According to the method, the brightness and noise of the low-illumination image are optimized through the attention mechanism and the capsule network, the image with enhanced brightness and clear details is obtained, and the image effect in the classified weak light environment is better.

Description

technical field [0001] The present invention relates to the technical fields of deep learning, image enhancement and image classification, in particular to a low-illumination image classification method based on an attention mechanism and a capsule network. Background technique [0002] In real life, there are a large number of images taken in low-light environments, and a lot of information cannot be accurately obtained during use, and it is difficult to distinguish with the naked eye, which leads to difficulties in image processing such as image classification, target detection, and restoration. While one major breakthrough after another has been achieved in image classification, most of them are dealing with bright images, and low-light processing is conspicuously lacking. This has always been a problem to be solved in the field of image classification technology. Contents of the invention [0003] Aiming at the deficiencies of the prior art, the present invention prov...

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): G06K9/62G06N3/04G06N3/08
CPCG06N3/084G06N3/047G06N3/048G06N3/045G06F18/2415G06F18/241
Inventor 江泽涛沈世琪
Owner GUILIN UNIV OF ELECTRONIC TECH
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