Attention based Covid-19 Detection using Generative Adversarial Network

Aiman Siddiqui, Asim Ahmed, Ali Faisal Saleem, Zeshan Khan Alvi, Tanvir Alam, Rizwan Qureshi

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

4 Citations (Scopus)

Abstract

The novel Coronavirus Disease 2019 (nCOVID-19) pandemic is a global health challenge, that requires collaborative efforts from multiple research communities. Effective screening of infected patients is a significant step in the fight against COVID-19, as radiological examination being an important screening methods. Early findings reveal that anomalies in chest X-rays of COVID-19 patients exist. As a result, a number of deep learning methods have been developed, and studies have shown that the accuracy of COVID-19 patient recognition using chest X-rays is very high. In this paper, we propose an attention based deep neural network for classifying the COVID-19 images, and extracting useful clinical information. Generative adversarial network is used to generate the synthetic COVID-19 images, as well as a good latent representation of both COVID-19 and normal images. Experiment results on public datasets shows the effectiveness of the proposed approach.

Original languageEnglish
Title of host publicationProceedings - 2021 IEEE 4th International Conference on Computing and Information Sciences, ICCIS 2021
EditorsMuhammad Taha Jilani
PublisherInstitute of Electrical and Electronics Engineers Inc.
ISBN (Electronic)9781665494410
DOIs
Publication statusPublished - 2021
Event4th IEEE International Conference on Computing and Information Sciences, ICCIS 2021 - Karachi, Pakistan
Duration: 29 Nov 202130 Nov 2021

Publication series

NameProceedings - 2021 IEEE 4th International Conference on Computing and Information Sciences, ICCIS 2021

Conference

Conference4th IEEE International Conference on Computing and Information Sciences, ICCIS 2021
Country/TerritoryPakistan
CityKarachi
Period29/11/2130/11/21

Keywords

  • Covid
  • Discriminator
  • GANs
  • Generator
  • Keras
  • MobileNet
  • ResNet

Fingerprint

Dive into the research topics of 'Attention based Covid-19 Detection using Generative Adversarial Network'. Together they form a unique fingerprint.

Cite this