Improved 3D U-Net for COVID-19 Chest CT Image Segmentation

Coronavirus disease 2019 (COVID-19) has spread rapidly worldwide. The rapid and accurate automatic segmentation of COVID-19 infected areas using chest computed tomography (CT) scans is critical for assessing disease progression. However, infected areas have irregular sizes and shapes. Furthermore, t...

Full description

Saved in:
Bibliographic Details
Published in:Scientific programming 2021, Vol.2021, p.1-9
Main Authors: Zheng, Ruiyong, Zheng, Yongguo, Dong-Ye, Changlei
Format: Article
Language:English
Subjects:
Artificial neural networks
Chest
Coders
Computed tomography
Coronaviruses
COVID-19
Datasets
Deep learning
Embedded systems
Feature extraction
Image segmentation
Infections
Lung diseases
Medical imaging
Neural networks
Pneumonia
Semantics
Teaching methods
Viral diseases
Citations: Items that this one cites
Items that cite this one
Online Access:Get full text
Tags: Add Tag
No Tags, Be the first to tag this record!
Description
Summary:Coronavirus disease 2019 (COVID-19) has spread rapidly worldwide. The rapid and accurate automatic segmentation of COVID-19 infected areas using chest computed tomography (CT) scans is critical for assessing disease progression. However, infected areas have irregular sizes and shapes. Furthermore, there are large differences between image features. We propose a convolutional neural network, named 3D CU-Net, to automatically identify COVID-19 infected areas from 3D chest CT images by extracting rich features and fusing multiscale global information. 3D CU-Net is based on the architecture of 3D U-Net. We propose an attention mechanism for 3D CU-Net to achieve local cross-channel information interaction in an encoder to enhance different levels of the feature representation. At the end of the encoder, we design a pyramid fusion module with expanded convolutions to fuse multiscale context information from high-level features. The Tversky loss is used to resolve the problems of the irregular size and uneven distribution of lesions. Experimental results show that 3D CU-Net achieves excellent segmentation performance, with Dice similarity coefficients of 96.3% and 77.8% in the lung and COVID-19 infected areas, respectively. 3D CU-Net has high potential to be used for diagnosing COVID-19.
ISSN:1058-9244
1875-919X
DOI:10.1155/2021/9999368