Super Resolution Reconstruction of MiniLED Based on Generative Adversarial Network

  • Yican Huang
  • , Xiaopin Zhong
  • , Weixiang Liu
  • , Zong Ze Wu

Research output: Chapter in Book/Report/Conference proceedingConference contributionpeer-review

Abstract

With the development of industrial detection technology, image super-resolution reconstruction is used as an effective data enhancement technique to provide the possibility of higher detection accuracy. This paper focuses on the image super-resolution in industrial detection scenarios, with the following two main contributions: 1) Constructing our own dataset in industrial detection specific scenarios. 2) Proposing EASRGAN, an image super-resolution reconstruction network, to effectively increase the accuracy of subsequent detection. Experimental results show that EASRGAN achieves excellent image super-resolution reconstruction, and the reconstructed image can increase the accuracy of the classification network by 2.83%.

Original languageEnglish
Title of host publicationProceedings - 2023 China Automation Congress, CAC 2023
PublisherInstitute of Electrical and Electronics Engineers Inc.
Pages7098-7103
Number of pages6
ISBN (Electronic)9798350303759
DOIs
StatePublished - 2023
Externally publishedYes
Event2023 China Automation Congress, CAC 2023 - Chongqing, China
Duration: 17 Nov 202319 Nov 2023

Publication series

NameProceedings - 2023 China Automation Congress, CAC 2023

Conference

Conference2023 China Automation Congress, CAC 2023
Country/TerritoryChina
CityChongqing
Period17/11/2319/11/23

Keywords

  • Generative adversarial network
  • Industrial detection
  • Super-resolution reconstruction
  • miniLED

Fingerprint

Dive into the research topics of 'Super Resolution Reconstruction of MiniLED Based on Generative Adversarial Network'. Together they form a unique fingerprint.

Cite this