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

A vehicle window image detail enhancement method and device based on a progressive optimization network

A progressive optimization, image technology, applied in the field of image super-resolution based on deep learning, can solve the problems of unable to reconstruct image details, noise in the collected data, etc., to achieve easy upsampling, easy large-scale enlargement, and improve computational efficiency Effect

Inactive Publication Date: 2019-06-28
HEFEI UNIV OF TECH
View PDF2 Cites 0 Cited by
  • Summary
  • Abstract
  • Description
  • Claims
  • Application Information

AI Technical Summary

Problems solved by technology

However, due to factors such as illumination changes, target occlusion, distance, and noise in the collected data, the traditional simple interpolation framework cannot reconstruct a finer image detail.
At the same time, for a super-resolution network based on deep learning, it is still very challenging to obtain a large-scale enlarged high-resolution image

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
  • A vehicle window image detail enhancement method and device based on a progressive optimization network
  • A vehicle window image detail enhancement method and device based on a progressive optimization network
  • A vehicle window image detail enhancement method and device based on a progressive optimization network

Examples

Experimental program
Comparison scheme
Effect test

Embodiment Construction

[0049]In this embodiment, a method for enhancing car window image details based on a progressive optimization network is to gradually reconstruct and enlarge the details in the traffic monitoring car window, including human faces, limbs, tissue boxes and decorations in the car, through a progressive optimization network To enhance its details, specifically, proceed as follows:

[0050] Step 1: Obtain m high-definition car window images from the database X={x k |k=1,2,…m}, where, x k Indicates the kth high-definition car window image;

[0051] For the kth high-definition car window image x k Perform n downsampling processing to obtain n downsampled images of the kth high-definition window image Thus, a set of n downsampled images of m high-definition car window images is obtained in, Indicates the kth high-definition window image x k After n-j times of downsampling, the kth high-definition window image x is obtained k The down-sampled car window image of the j-th laye...

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 vehicle window image detail enhancement method and device based on a progressive optimization network. The method comprises the following steps that reconstruction and amplification are performed on a small low-resolution vehicle window image for multiple times, then the mean square error between the image reconstructed and amplified each time and a corresponding downsampling image is calculated, and a progressive optimization network is trained by utilizing a random gradient descent method or a batch gradient descent method, so that the vehicle window image in a traffic monitoring video is reconstructed into a high-definition vehicle window image. According to the invention, the performance of the super-resolution network based on deep learning can be improved, so that the vehicle window image obtained in the monitoring video can be reconstructed into a super-large-scale amplified high-quality clear vehicle window image.

Description

technical field [0001] The invention relates to the technical field of image super-resolution based on deep learning, in particular to a method and device for enhancing details of car window images based on progressive optimization network. Background technique [0002] Image super-resolution, that is, super-analyzing a low-resolution image to a high-resolution image, is a hot research issue in the field of computer vision and graphics, and deep learning is also the most promising direction in the field of machine learning. Its image super-resolution has a wide range of applications in surveillance video security, face recognition, face tracking, facial expression estimation, etc. However, due to factors such as illumination changes, target occlusion, distance, and noise in the collected data, the traditional simple interpolation framework cannot reconstruct a finer image detail. At the same time, for a super-resolution network based on deep learning, it is still very chall...

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): G06T5/50
Inventor 赵洋刘小娜贾伟陈缘李书杰李琳刘晓平
Owner HEFEI UNIV OF TECH
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