Related Experiment Video
Updated: Sep 26, 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
14.3K
A computer-aided diagnostic framework for coronavirus diagnosis using texture-based radiomics images.
1Department of Electronics and Communications Engineering, College of Engineering and Technology, Arab Academy for Science, Technology and Maritime Transport, Alexandria, Egypt.
Digital Health
|April 18, 2022
Summary
This study introduces a novel deep learning framework using texture-based radiomics for improved coronavirus detection in CT scans. Combining multiple radiomics features significantly enhances diagnostic accuracy and performance for rapid disease identification.
Area of Science:
- Medical Imaging
- Artificial Intelligence
- Radiology
Background:
- Accurate and rapid detection of coronavirus is crucial for disease containment and mitigating healthcare impacts.
- Artificial intelligence (AI), specifically deep learning, shows promise for precise coronavirus diagnosis from computed tomography (CT) images.
- Integrating texture-based radiomics with deep learning can potentially improve diagnostic accuracy compared to using original CT images.
Purpose of the Study:
- To propose and validate a computer-assisted diagnostic framework for coronavirus detection.
- To enhance diagnostic accuracy by integrating texture-based radiomics (discrete wavelet transform, gray-level covariance matrix) with deep learning models (ResNets).
- To evaluate the performance of fused radiomic features from multiple deep learning networks.
Main Methods:
- Three Residual Networks (ResNets) were trained using discrete wavelet transform and gray-level covariance matrix radiomics images instead of original CT images.
- Texture-based radiomics deep features were extracted, fused using discrete cosine transform, and further combined from the three ResNets.
- Support vector machine classifiers were employed for the final classification of severe respiratory syndrome coronavirus 2 (SARS-CoV-2) CT images.
Main Results:
- Texture-based radiomics images significantly improved ResNet performance (e.g., ResNet-18: 83.22% vs. 70.34%) compared to original CT images.
- The proposed framework achieved high diagnostic performance with sensitivity, specificity, accuracy, precision, and F1-score all exceeding 99% after feature fusion.
- Combining multiple texture-based radiomics deep features from several deep learning models outperformed single approaches.
Conclusions:
- The integration of texture-based radiomics with deep learning, particularly through feature fusion from multiple networks, substantially boosts coronavirus diagnostic accuracy.
- The developed computer-assisted diagnostic framework demonstrates high efficacy and can be a valuable tool for radiologists in achieving fast and accurate diagnoses.
- This approach offers a promising advancement in AI-driven medical diagnostics for infectious diseases like coronavirus.
