Related Experiment Video
Updated: Nov 6, 2025
Brain Infarct Segmentation and Registration on MRI or CT for Lesion-symptom Mapping
Published on: September 25, 2019
COVID-19 lesion discrimination and localization network based on multi-receptive field attention module on CT images
Xia Ma1,2, Bingbing Zheng3, Yu Zhu3
1Department of Pulmonary and Critical Care Medicine, The Third Hospital of Shanxi Medical University, Taiyuan 030032, China.
Insights
A new deep learning network with a multi-receptive field attention module aids in diagnosing COVID-19 from CT scans. This AI tool improves accuracy and localization, assisting doctors in identifying the disease effectively.
Area of Science:
- Medical Imaging
- Artificial Intelligence
- Computer-Aided Diagnosis
Background:
- Coronavirus Disease 2019 (COVID-19) remains a global health concern since its emergence in late 2019.
- While RT-PCR is the gold standard, Computed Tomography (CT) imaging is crucial for COVID-19 diagnosis and treatment evaluation.
- Deep learning offers potential for quantitative analysis and improved diagnostic accuracy on CT images.
Purpose of the Study:
- To develop and evaluate a novel deep learning network for diagnosing COVID-19 on CT images.
- To enhance diagnostic capabilities by incorporating a multi-receptive field attention module.
- To improve the localization and discrimination of COVID-19 lesions.
Main Methods:
- A novel deep learning network incorporating a multi-receptive field attention module was proposed.
- The attention module consists of a Pyramid Convolutional Module (PCM), Spatial Attention Block (SAB), and Channel Attention Block (CAB).
- The method was validated on two distinct datasets, including one from Beijing Ditan Hospital.
Main Results:
- The proposed network achieved high performance metrics: 97.12% accuracy, 96.89% specificity, and 97.21% sensitivity on the DTDB dataset.
- On a public COVID-19 dataset, the network obtained 95.16% accuracy, 95.6% F1-score, and 99.01% AUC.
- The multi-receptive field attention module demonstrated superior performance compared to other state-of-the-art attention mechanisms.
Conclusions:
- The developed deep learning network with a multi-receptive field attention module effectively diagnoses COVID-19 from CT images.
- The network provides valuable quantitative auxiliary information for clinicians, aiding in diagnosis and localization.
- This AI-driven approach shows significant promise in assisting medical professionals in the fight against COVID-19.
Abstract:
Since discovered in Hubei, China in December 2019, Corona Virus Disease 2019 named COVID-19 has lasted more than one year, and the number of new confirmed cases and confirmed deaths is still at a high level. COVID-19 is an infectious disease caused by SARS-CoV-2. Although RT-PCR is considered the gold standard for detection of COVID-19, CT plays an important role in the diagnosis and evaluation of the therapeutic effect of COVID-19. Diagnosis and localization of COVID-19 on CT images using deep learning can provide quantitative auxiliary information for doctors. This article proposes a novel network with multi-receptive field attention module to diagnose COVID-19 on CT images. This attention module includes three parts, a pyramid convolution module (PCM), a multi-receptive field spatial attention block (SAB), and a multi-receptive field channel attention block (CAB). The PCM can improve the diagnostic ability of the network for lesions of different sizes and shapes. The role of SAB and CAB is to focus the features extracted from the network on the lesion area to improve the ability of COVID-19 discrimination and localization. We verify the effectiveness of the proposed method on two datasets. The accuracy rate of 97.12%, specificity of 96.89%, and sensitivity of 97.21% are achieved by the proposed network on DTDB dataset provided by the Beijing Ditan Hospital Capital Medical University. Compared with other state-of-the-art attention modules, the proposed method achieves better result. As for the public COVID-19 SARS-CoV-2 dataset, 95.16% for accuracy, 95.6% for F1-score and 99.01% for AUC are obtained. The proposed network can effectively assist doctors in the diagnosis of COVID-19 CT images.
