City Research Online

Abstract

Abstract—We introduce a lightweight model derived from Mask R-CNN that segments lesions and predicts COVID-19 from chest CT scans in a single shot. The model requires a small dataset to train, and is evaluated on a large set of images to achieve a 42.45% average precision on the segmentation test split, and 93.00% COVID-19 sensitivity and F1-score of 96.76% on the classification test split across 3 classes: COVID-19, Common Pneumonia and Negative. We introduce an augmented Region of Interest layer that disentangles lesion detection functionality for segmentation and classification problems. Efficiency of the solution is confirmed by comparing it to a suite of the state-of-the-art models across both problems. Full source code, models and pretrained weights are available on https://github.com/AlexTS1980/COVID-Single-Shot-Model.

Publication Type:	Monograph (Working Paper)
Publisher Keywords:	COVID-19, Instance Segmentation, Object Detection, Regions of Interest, Image Classification
Subjects:	H Social Sciences > HN Social history and conditions. Social problems. Social reform Q Science > QA Mathematics > QA75 Electronic computers. Computer science Q Science > QR Microbiology R Medicine > RA Public aspects of medicine > RA0421 Public health. Hygiene. Preventive Medicine R Medicine > RC Internal medicine
Departments:	School of Science & Technology > Computer Science

Export

Downloads