Eureka AIR delivers breakthrough ideas for toughest innovation challenges, trusted by R&D personnel around the world.

Image defogging and rain-removing algorithm based on selective attention mechanism

An attention and selective technology, applied in the field of image restoration, can solve the problem of insufficient restoration effect, and achieve the effect of accurate rain/fog removal

Inactive Publication Date: 2020-09-11
NANJING UNIV OF SCI & TECH
View PDF0 Cites 13 Cited by
  • Summary
  • Abstract
  • Description
  • Claims
  • Application Information

AI Technical Summary

Problems solved by technology

[0004] In view of the existing image defogging and rain removal methods, which rely on the physical model and the restoration effect is not good enough, the present invention provides an image defogging and rain removal algorithm based on a selective attention mechanism.

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 defogging and rain-removing algorithm based on selective attention mechanism
  • Image defogging and rain-removing algorithm based on selective attention mechanism
  • Image defogging and rain-removing algorithm based on selective attention mechanism

Examples

Experimental program
Comparison scheme
Effect test

Embodiment

[0053] The present invention proposes an image defogging and deraining algorithm based on a selective attention mechanism, including expanding training samples, training the network on the training set, defogging and deraining the test pictures, and combining the restored image with the clear image Three processes are assessed:

[0054] Expanding the training samples includes the following steps:

[0055] Step 1: The RESIDE indoor data set is used for the fog removal data set, and the Rain100H and Rain100L data sets are used for the rain removal data set. A part of the pictures in the training set in the data set are used as the verification set, and the remaining pictures in the training set are cropped, intercepted, rotated, etc. The original data set is expanded so that the image sizes in the defog and derain data sets are 640*460 and 100*100 respectively, which can avoid overfitting during training.

[0056] Training the network on the training set involves the following ...

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 defogging and rain-removing algorithm based on a selective attention mechanism. The image defogging and rain-removing algorithm specifically comprises the following three processes of expanding a training sample, training a network on a training set, defogging and removing rain from a test image, and evaluating a restored image and a clear image. The method comprises s that: firstly, a defogging data set conducts cutting, intercepting, rotating and other operations on a used training set to expand an original data set; the pictures are input into a convolutional network based on a selective attention mechanism, and the network is trained by using a multi-loss function; and finally, the test pictures are input into the network to obtain defogging / rain-removing results. According to the invention, an end-to-end defogging / rain removing algorithm is realized, and rain-removing / defogging can be efficiently and accurately carried out on an image with rain / fog.

Description

technical field [0001] The invention relates to the technical field of image restoration, in particular to an image defogging and rain removal algorithm based on a selective attention mechanism. Background technique [0002] Rain and smog are becoming more and more common in our daily life. Images captured in hazy and rainy weather often have relatively low visibility and contrast. However, various vision tasks, such as image classification, semantic segmentation, and object detection, require sharp images as input. Smog and rain bands make these tasks impossible to complete well. Since the direction and density of rain bands in rainfall images and the depth of fog and atmospheric illumination in foggy images are inhomogeneous, image dehazing and denoising are ill-posed problems. In view of this, the research on image defogging and rain removal is of great significance. [0003] Most of the traditional fog and rain removal algorithms restore the final clear image based o...

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): G06T5/00G06N3/04G06N3/08G06K9/62
CPCG06N3/08G06N3/045G06F18/214G06T5/73
Inventor 唐金辉梁潇
Owner NANJING UNIV OF SCI & 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
Eureka Blog
Learn More
PatSnap group products