City Research Online

Abstract

We present a model that fuses lesion segmentation with Attention Mechanism to predict COVID-19 from chest CT scans. The model segments lesions, extracts Regions of Interest from scans and applies Attention to them to determine the most relevant ones for image classification. Additionally, we augment the model with Long-Short Term Memory Network layers that learn features from a sequence of Regions of Interest before computing attention. The model is trained in one shot for both problems, using two different sets of data. We achieve 0.4683 mean average precision for lesion segmentation, 95.74% COVID-19 sensitivity and 98.15% class-adjusted F1 score for image classification on a large CNCB-NCOV dataset. Source code is available on https: //github.com/AlexTS1980/COVID-LSTM-Attention.

Publication Type:	Article
Additional Information:	© 2021 IEEE. Personal use of this material is permitted. Permission from IEEE must be obtained for all other uses, in any current or future media, including reprinting/republishing this material for advertising or promotional purposes, creating new collective works, for resale or redistribution to servers or lists, or reuse of any copyrighted component of this work in other works.
Subjects:	H Social Sciences > HN Social history and conditions. Social problems. Social reform Q Science > QA Mathematics > QA76 Computer software Q Science > QR Microbiology R Medicine > RA Public aspects of medicine > RA0421 Public health. Hygiene. Preventive Medicine
Departments:	School of Science & Technology > Computer Science

Export

Downloads