Related Experiment Video
Updated: Nov 19, 2025
08:05
Lung CT Segmentation to Identify Consolidations and Ground Glass Areas for Quantitative Assesment of SARS-CoV Pneumonia
Published on: December 19, 2020
14.5K
LungINFseg: Segmenting COVID-19 Infected Regions in Lung CT Images Based on a Receptive-Field-Aware Deep Learning
Vivek Kumar Singh1, Mohamed Abdel-Nasser1,2, Nidhi Pandey3
1Department of Computer Engineering and Mathematics, Universitat Rovira i Virgili, 43007 Tarragona, Spain.
Diagnostics (Basel, Switzerland)
|January 27, 2021
Summary
A new deep learning method, LungINFseg, accurately segments COVID-19 infections in lung CT scans. This automated approach improves early disease identification, crucial for managing the pandemic when hospital resources are limited.
Area of Science:
- Medical Imaging
- Artificial Intelligence in Healthcare
- Radiology
Background:
- COVID-19 pandemic necessitates rapid and accurate patient identification.
- Lung computed tomography (CT) imaging serves as a vital diagnostic tool.
- Manual segmentation of lung CT images is labor-intensive and prone to inaccuracies.
Purpose of the Study:
- To develop a fully automated and efficient deep learning method for segmenting COVID-19 infections in lung CT images.
- To address limitations of existing segmentation methods, including texture variations and pathological changes.
Main Methods:
- Introduction of LungINFseg, a novel deep learning-based segmentation technique.
- Development of the receptive-field-aware (RFA) module to enhance feature learning.
- RFA module integrates convolution layers, dilated convolution, discrete wavelet transform, and an attention mechanism.
Main Results:
- LungINFseg demonstrated superior performance on a dataset of over 1800 annotated CT slices.
- Achieved a Dice score of 80.34% and an Intersection-over-Union (IoU) score of 68.77%.
- Outperformed 13 state-of-the-art deep learning segmentation methods, including U-Net, by approximately 10% in Dice and IoU scores.
Conclusions:
- LungINFseg offers an effective and automated solution for COVID-19 infection segmentation in lung CT images.
- The RFA module's ability to capture contextual information improves segmentation accuracy.
- This method holds potential for aiding in early COVID-19 diagnosis and disease management.
