Related Experiment Video
Updated: Oct 29, 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
14.4K
COVSeg-NET: A deep convolution neural network for COVID-19 lung CT image segmentation
XiaoQing Zhang1, GuangYu Wang2, Shu-Guang Zhao2
1Taizhou Institute of Science and Technology, Nanjing University of Science and Technology No.8, Meilan East Road Taizhou China.
Summary
A new AI model, COVSeg-NET, accurately segments COVID-19 lung lesions from CT scans. This artificial intelligence approach offers faster, more effective detection of ground glass opaque lesions, improving patient diagnosis.
Area of Science:
- Medical Imaging
- Artificial Intelligence
- Computer-Aided Diagnosis
Background:
- COVID-19 is a global respiratory pandemic requiring rapid diagnostic tools.
- Accurate segmentation of lung lesions in CT images is crucial for COVID-19 assessment.
- Existing methods may lack efficiency or accuracy in lesion detection.
Purpose of the Study:
- To develop and evaluate a novel AI model, COVSeg-NET, for precise segmentation of COVID-19 related ground glass opaque lesions.
- To demonstrate the efficacy of COVSeg-NET in analyzing lung CT images.
Main Methods:
- The COVSeg-NET model was designed using a fully convolutional neural network architecture.
- Key components include convolutional layers, activation functions, pooling, batch normalization, and sigmoid layers.
- The model was trained and tested on COVID-19 lung CT images.
Main Results:
- COVSeg-NET achieved a Dice coefficient of 0.561, sensitivity of 0.447, and specificity of 0.996.
- Performance metrics indicate superior results compared to other deep learning methods.
- The model demonstrated effectiveness with smaller training datasets and reduced testing times.
Conclusions:
- COVSeg-NET provides an advanced and accurate method for segmenting COVID-19 lung lesions.
- The model's efficiency and accuracy support its potential for rapid and effective COVID-19 detection.
- This AI-driven approach offers a promising tool for medical image analysis in pandemic scenarios.
