Dense Regression Activation Maps For Lesion Segmentation in CT scans of COVID-19 patients

TL;DR

This paper introduces a weakly-supervised method using dense regression activation maps and an attention module for accurate lesion segmentation in COVID-19 CT scans, reducing annotation costs and improving segmentation accuracy.

Contribution

The authors propose a novel dense regression activation map approach combined with an attention module for lesion segmentation, outperforming traditional CAM-based methods.

Findings

Achieved 70.2% Dice coefficient on COVID-19 CT scans.

Outperformed CAM-based baseline with 48.6% Dice coefficient.

Demonstrated effectiveness with 90 subjects.

Abstract

Automatic lesion segmentation on thoracic CT enables rapid quantitative analysis of lung involvement in COVID-19 infections. However, obtaining a large amount of voxel-level annotations for training segmentation networks is prohibitively expensive. Therefore, we propose a weakly-supervised segmentation method based on dense regression activation maps (dRAMs). Most weakly-supervised segmentation approaches exploit class activation maps (CAMs) to localize objects. However, because CAMs were trained for classification, they do not align precisely with the object segmentations. Instead, we produce high-resolution activation maps using dense features from a segmentation network that was trained to estimate a per-lobe lesion percentage. In this way, the network can exploit knowledge regarding the required lesion volume. In addition, we propose an attention neural network module to refine dRAMs, optimized together with the main regression task. We evaluated our algorithm on 90 subjects. Results show our method achieved 70.2% Dice coefficient, substantially outperforming the CAM-based baseline at 48.6%.