Underwater image real-time restoration model based on self-attention mechanism and GAN

An underwater image and attention technology, which is applied in the field of image processing and deep learning, can solve problems such as blurring, limited visible range of underwater images, and color incongruity, and achieve diverse scenes, real-time improvement of visual quality, and superior restoration performance effect

Active Publication Date: 2021-08-03
GUILIN UNIV OF ELECTRONIC TECH +1
View PDF16 Cites 2 Cited by
  • Summary
  • Abstract
  • Description
  • Claims
  • Application Information

AI Technical Summary

Problems solved by technology

The complexity of the marine environment is diverse, and multiple unfavorable factors such as light transmission in water will be seriously attenuated by water absorption, reflection,

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
  • Underwater image real-time restoration model based on self-attention mechanism and GAN
  • Underwater image real-time restoration model based on self-attention mechanism and GAN
  • Underwater image real-time restoration model based on self-attention mechanism and GAN

Examples

Experimental program
Comparison scheme
Effect test

Embodiment Construction

[0045] A real-time restoration model of underwater images based on the self-attention mechanism and GAN of the present invention will be described in further detail below in conjunction with specific embodiments.

[0046] The real-time restoration model of underwater images based on the self-attention mechanism and GAN proposed by the present invention introduces a self-attention module, and the specific implementation steps are as follows:

[0047] S1. Related work:

[0048] S1.1. Generative Adversarial Network Model: GAN is a deep neural network composed of two networks, a generator and a discriminator. Both the generator and the discriminator adopt the idea of ​​zero-sum game in game theory, in which the goal of the generator is to learn the mapping relationship between the degraded underwater images and the clear underwater images, so as to generate high-quality clear underwater images; The goal is to make the network learn to distinguish generated fake images from real r...

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 relates to an underwater image real-time restoration model based on a self-attention mechanism and a GAN, and belongs to the technical field of deep learning. A generative adversarial network is used as a basic framework in the invention, the generative network adopts a coding-decoding structure, a synthetic image is generated at an original resolution through nine residual blocks and deconvolution operation, a self-attention module is introduced, richer advanced features can be captured to improve model performance, and in order to keep image content and remove underwater noise at the same time, the discrimination network adopts a structure of a multi-branch discriminator comprising an adversarial branch and a criticality branch. The self-attention module is added, the model structure is improved, the problem of low time efficiency of underwater image processing is solved ideally, and the method based on GAN-RS is trained to adapt to various underwater conditions, and has excellent real-time processing performance.

Description

technical field [0001] The invention relates to the technical fields of image processing and deep learning, and more specifically, the invention is oriented to the restoration task of underwater degraded images. Background technique [0002] In recent years, underwater robots have been widely used in marine resource exploration and other aspects. The enhancement and restoration technology of underwater degraded images has attracted much attention due to its significance for marine exploration, development and utilization. The complexity of the marine environment is diverse, and multiple unfavorable factors such as light transmission in water will be seriously attenuated by water absorption, reflection, and scattering. The collected underwater images will inevitably have limited visible range, blurred, Issues such as low contrast, non-uniform lighting, color incongruities, and noise. Underwater image enhancement and restoration technology aims to repair the degraded underwat...

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
CPCG06T5/002G06N3/084G06N3/045
Inventor 罗笑南刘瑞关善文秦子钦孙钦李冀吕智张敏
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