Related Experiment Video
Updated: Dec 10, 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
14.6K
COVID-19 Deep Learning Prediction Model Using Publicly Available Radiologist-Adjudicated Chest X-Ray Images as
Mohd Zulfaezal Che Azemin1, Radhiana Hassan2, Mohd Izzuddin Mohd Tamrin3
1Kulliyyah of Allied Health Sciences, International Islamic University Malaysia, Bandar Indera Mahkota, 25200 Kuantan, Pahang, Malaysia.
International Journal of Biomedical Imaging
|August 28, 2020
Summary
This study developed a deep learning model for detecting COVID-19 from chest X-rays using clinically associated images for training. The model achieved promising results, addressing data limitations in deep learning for COVID-19 detection.
Area of Science:
- Artificial Intelligence in Medical Imaging
- Deep Learning for Disease Detection
- Radiology and Pulmonary Medicine
Background:
- Limited availability of public COVID-19 chest X-ray datasets hinders deep learning model generalization.
- Existing models may lack robustness due to small, potentially biased training data.
- Need for reliable AI tools to aid in COVID-19 diagnosis using medical imaging.
Purpose of the Study:
- To develop and evaluate a deep learning model for COVID-19 detection using chest radiographs.
- To overcome data scarcity by utilizing clinically associated findings for training.
- To ensure model robustness by using mutually exclusive training and testing datasets.
Main Methods:
- Utilized a large dataset of chest radiograph images with clinical findings suggestive of COVID-19 for training.
- Employed a ResNet-101 convolutional neural network architecture, pre-trained on a large object recognition dataset.
- Retrained the model specifically for detecting abnormalities in chest X-ray images and validated on confirmed COVID-19 cases.
Main Results:
- The deep learning model achieved an Area Under the Receiver Operating Curve (AUC) of 0.82.
- The model demonstrated a sensitivity of 77.3%, specificity of 71.8%, and accuracy of 71.9%.
- The study highlights the effectiveness of using clinically associated labels and mutually exclusive datasets.
Conclusions:
- Deep learning models can be effectively trained for COVID-19 detection using chest X-rays with strong clinical associations.
- The approach of using clinically labeled data for training and confirmed cases for testing enhances model reliability.
- This method offers a viable solution to the challenge of limited, confirmed COVID-19 imaging data for AI development.
