A Novel Deep Learning Algorithm for Covid Detection and Classification

S. Selvi, Nikhil Agarwal, Paarth Barkur, Yash Mishra, Abhsihek Kumar
{"title":"A Novel Deep Learning Algorithm for Covid Detection and Classification","authors":"S. Selvi, Nikhil Agarwal, Paarth Barkur, Yash Mishra, Abhsihek Kumar","doi":"10.1109/CONIT55038.2022.9847880","DOIUrl":null,"url":null,"abstract":"The prediction of future development of a natural phenomenon is one of the main objectives of recent technology, but this is a great challenge when dealing with an epidemic or pandemic. This proved to be particularly true in the case of Covid-19 global pandemic that the world is suffering and facing since January 2020. The response to the virus infection are partially known, however the immune system is mostly affected especially in patients with pre-existing respiratory or systemic diseases. Most infections by coronavirus are mild and self-treated. Therefore, in early stages of the disease, it will be misleading to estimate the real spread of the virus based on the reports of hospital. Moreover, such reports vary according to how measurements are performed, and the number of tests related only to the number of symptomatic patients. Despite all this, the large amount of official data published in last months, and updated daily has motivated various mathematical models, which are required to predict the evolution of an epidemic and plan effective control strategies. Due to the incompleteness of the data and intrinsic complexity, predicting the evolution, the peak or the end of the pandemic is a challenge. In this paper, a deep learning based approach is proposed aiming to evaluate a-priori risk of an epidemic caused by Covid-19. The proposed algorithm leverages image processing and deep learning algorithms to detect Covid and differentiate between normal, Covid affected, lung opacity and viral pneumonia affected chest x-rays. This results in setting strategies to prevent or decrease the impact of future epidemic waves. The accuracy for the proposed algorithm is 95.01% and Recall is 98.5% on validation data. The inference is that combining image processing with deep learning can improve performance of Covid detection.","PeriodicalId":270445,"journal":{"name":"2022 2nd International Conference on Intelligent Technologies (CONIT)","volume":"24 1","pages":"0"},"PeriodicalIF":0.0000,"publicationDate":"2022-06-24","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":"0","resultStr":null,"platform":"Semanticscholar","paperid":null,"PeriodicalName":"2022 2nd International Conference on Intelligent Technologies (CONIT)","FirstCategoryId":"1085","ListUrlMain":"https://doi.org/10.1109/CONIT55038.2022.9847880","RegionNum":0,"RegionCategory":null,"ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":null,"EPubDate":"","PubModel":"","JCR":"","JCRName":"","Score":null,"Total":0}
引用次数: 0

Abstract

The prediction of future development of a natural phenomenon is one of the main objectives of recent technology, but this is a great challenge when dealing with an epidemic or pandemic. This proved to be particularly true in the case of Covid-19 global pandemic that the world is suffering and facing since January 2020. The response to the virus infection are partially known, however the immune system is mostly affected especially in patients with pre-existing respiratory or systemic diseases. Most infections by coronavirus are mild and self-treated. Therefore, in early stages of the disease, it will be misleading to estimate the real spread of the virus based on the reports of hospital. Moreover, such reports vary according to how measurements are performed, and the number of tests related only to the number of symptomatic patients. Despite all this, the large amount of official data published in last months, and updated daily has motivated various mathematical models, which are required to predict the evolution of an epidemic and plan effective control strategies. Due to the incompleteness of the data and intrinsic complexity, predicting the evolution, the peak or the end of the pandemic is a challenge. In this paper, a deep learning based approach is proposed aiming to evaluate a-priori risk of an epidemic caused by Covid-19. The proposed algorithm leverages image processing and deep learning algorithms to detect Covid and differentiate between normal, Covid affected, lung opacity and viral pneumonia affected chest x-rays. This results in setting strategies to prevent or decrease the impact of future epidemic waves. The accuracy for the proposed algorithm is 95.01% and Recall is 98.5% on validation data. The inference is that combining image processing with deep learning can improve performance of Covid detection.
一种新型的Covid检测与分类深度学习算法
预测自然现象的未来发展是现代技术的主要目标之一,但在处理流行病或大流行病时,这是一个巨大的挑战。事实证明,自2020年1月以来,世界正在遭受和面临的Covid-19全球大流行尤其如此。对病毒感染的反应尚不完全清楚,但免疫系统主要受到影响,特别是在已有呼吸道或全身性疾病的患者中。大多数冠状病毒感染是轻微的,可以自我治疗。因此,在疾病的早期阶段,根据医院的报告估计病毒的真实传播情况会产生误导。此外,这些报告因测量方式的不同而不同,检测次数仅与有症状患者的数量有关。尽管如此,过去几个月公布的大量官方数据,以及每天更新的数据,催生了各种数学模型,这些模型是预测疫情演变和制定有效控制战略所必需的。由于数据的不完整和内在的复杂性,预测大流行的演变、高峰或结束是一项挑战。本文提出了一种基于深度学习的方法,旨在评估Covid-19引起的流行病的先验风险。该算法利用图像处理和深度学习算法来检测Covid,并区分正常、受Covid影响、肺不透明和病毒性肺炎影响的胸部x光片。这导致制定战略,以防止或减少未来流行病浪潮的影响。该算法在验证数据上的准确率为95.01%,召回率为98.5%。由此推断,将图像处理与深度学习相结合可以提高Covid检测的性能。
本文章由计算机程序翻译,如有差异,请以英文原文为准。
求助全文
约1分钟内获得全文 求助全文
来源期刊
自引率
0.00%
发文量
0
×
引用
GB/T 7714-2015
复制
MLA
复制
APA
复制
导出至
BibTeX EndNote RefMan NoteFirst NoteExpress
×
提示
您的信息不完整,为了账户安全,请先补充。
现在去补充
×
提示
您因"违规操作"
具体请查看互助需知
我知道了
×
提示
确定
请完成安全验证×
copy
已复制链接
快去分享给好友吧!
我知道了
右上角分享
点击右上角分享
0
联系我们:info@booksci.cn Book学术提供免费学术资源搜索服务,方便国内外学者检索中英文文献。致力于提供最便捷和优质的服务体验。 Copyright © 2023 布克学术 All rights reserved.
京ICP备2023020795号-1
ghs 京公网安备 11010802042870号
Book学术文献互助
Book学术文献互助群
群 号:481959085
Book学术官方微信