Dmbg-Net: Dilated multiresidual boundary guidance network for COVID-19 infection segmentation
Related Experiment Video
Updated: Jul 9, 2025
Lung CT Segmentation to Identify Consolidations and Ground Glass Areas for Quantitative Assesment of SARS-CoV Pneumonia
Published on: December 19, 2020
Zhenwu Xiang1, Qi Mao1, Jintao Wang1
1School of Electronic and Electrical Engineering, Shanghai University of Engineering Science, Shanghai 201620, China.
Insights
A new deep learning model, Dmbg-Net, accurately segments COVID-19 lung infections in CT scans. This method enhances boundary detection and feature representation for improved COVID-19 diagnosis.
Area of Science:
- Medical Imaging
- Artificial Intelligence
- Computer Vision
Background:
- Accurate segmentation of lung lesions in CT images is crucial for COVID-19 detection.
- Challenges include obscure boundaries, low contrast, and scattered infection areas.
Purpose of the Study:
- To propose a novel network, Dmbg-Net, for improved COVID-19 infection segmentation in lung CT images.
- To address segmentation challenges by focusing on semantic relationships and boundary guidance.
Main Methods:
- Utilized dilated residual blocks and dilated convolutions to expand receptive fields and minimize feature loss.
- Incorporated an edge-attention guidance preservation block for low-level feature boundary integration.
- Employed progressive multi-scale supervision for enhanced representations and accurate saliency maps.
Main Results:
- Achieved a Dice similarity coefficient of 85.6% and a sensitivity of 84.2% on COVID-19 datasets.
- Ablation studies confirmed the effectiveness of the Dmbg-Net architecture.
- Demonstrated potential for segmenting other types of lung lesions.
Conclusions:
- Dmbg-Net effectively segments COVID-19 infections in lung CT images.
- The method shows promise for clinical applications in lesion detection and segmentation.
Abstract:
Accurate segmentation of infected regions in lung computed tomography (CT) images is essential for the detection and diagnosis of coronavirus disease 2019 (COVID-19). However, lung lesion segmentation has some challenges, such as obscure boundaries, low contrast and scattered infection areas. In this paper, the dilated multiresidual boundary guidance network (Dmbg-Net) is proposed for COVID-19 infection segmentation in CT images of the lungs. This method focuses on semantic relationship modelling and boundary detail guidance. First, to effectively minimize the loss of significant features, a dilated residual block is substituted for a convolutional operation, and dilated convolutions are employed to expand the receptive field of the convolution kernel. Second, an edge-attention guidance preservation block is designed to incorporate boundary guidance of low-level features into feature integration, which is conducive to extracting the boundaries of the region of interest. Third, the various depths of features are used to generate the final prediction, and the utilization of a progressive multi-scale supervision strategy facilitates enhanced representations and highly accurate saliency maps. The proposed method is used to analyze COVID-19 datasets, and the experimental results reveal that the proposed method has a Dice similarity coefficient of 85.6% and a sensitivity of 84.2%. Extensive experimental results and ablation studies have shown the effectiveness of Dmbg-Net. Therefore, the proposed method has a potential application in the detection, labeling and segmentation of other lesion areas.
