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

Super-resolution implementation method based on reconstruction optimization and deep neural network

A deep neural network and super-resolution technology, which is applied in the field of super-resolution realization based on reconstruction optimization and deep neural network, can solve problems such as image information inconsistency and errors, and achieve excellent performance.

Active Publication Date: 2018-03-09
NANJING UNIV
View PDF6 Cites 17 Cited by
  • Summary
  • Abstract
  • Description
  • Claims
  • Application Information

AI Technical Summary

Problems solved by technology

However, despite this, these learning-based methods directly speculate on the lost high-frequency details based on the low-resolution input image and the learned mapping function, which will likely lead to erroneous results that are different from the real image information. does not match

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
  • Super-resolution implementation method based on reconstruction optimization and deep neural network
  • Super-resolution implementation method based on reconstruction optimization and deep neural network
  • Super-resolution implementation method based on reconstruction optimization and deep neural network

Examples

Experimental program
Comparison scheme
Effect test

Embodiment Construction

[0029] The present invention will be described in detail below in conjunction with the accompanying drawings and specific embodiments.

[0030] The present invention first adopts a reconstruction-based super-resolution technique, taking multiple low-resolution image sequences with sub-pixel offsets as input. Considering the ill-conditioned nature of the problem, the L1 norm is added to the objective function to constrain the solution process. The L1 regularization term helps to generate a sparse weight matrix, which in turn can be used for feature selection. The objective function is quickly converged by conjugate gradients to obtain a high-resolution image. In reconstruction-based super-resolution techniques, ringing inevitably occurs when the magnification factor is too large or the input image is insufficient. In the present invention, a three-layer fully convolutional neural network is used to suppress the ringing phenomenon, and further optimize the previously output hi...

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 super-resolution implementation method based on reconstruction optimization and a deep neural network. The method comprises the following specific steps of 1, establishing arelationship between a high-resolution image and a down-sampled image by means of a down-sampling method, and establishing a target function by means of a least square method; 2, utilizing a conjugategradient descent algorithm to conduct iterative optimization on the target function to obtain a high-resolution image based on a reconstruction optimization algorithm; 3, establishing and completinga three-layer full convolutional neural network through training; 4, inputting the high-resolution image obtained in the step 2 into the three-layer convolutional neural network to further improve theresolution of the image. By means of the method, good results can be achieved in both subjective evaluation and objective image quality evaluation.

Description

technical field [0001] The invention relates to the field of computational photography, in particular to a super-resolution implementation method based on reconstruction optimization and deep neural network. Background technique [0002] Image super-resolution techniques aim to recover high-resolution images with more high-frequency details from one (single-frame super-resolution) or series (multi-frame super-resolution) of low-resolution images. Existing algorithms for image super-resolution are mainly divided into two categories: reconstruction-based methods and learning-based methods. [0003] The reconstruction-based algorithm reconstructs a high-resolution image from a series of low-resolution images with sub-pixel offsets by simulating the inverse process of downsampling. However, this reconstruction-based method is essentially an ill-conditioned process because of the lack of high-frequency detail information of the image. This problem can be solved to a certain ext...

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): G06T3/40G06N3/04
CPCG06T3/4053G06T3/4069G06N3/045
Inventor 马展吴洁
Owner NANJING UNIV
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