Related Experiment Video
Updated: Oct 25, 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
14.4K
A novel deep learning based method for COVID-19 detection from CT image
SeyyedMohammad JavadiMoghaddam1, Hossain Gholamalinejad2
1Department of Computer Engineering, Bozorgmehr University of Qaenat, Qaen, Iran.
Biomedical Signal Processing and Control
|August 4, 2021
Summary
A novel deep learning model aids COVID-19 diagnosis using CT scans. This model achieves 99.03% accuracy, offering a fast and effective auxiliary detection tool for the pandemic.
Area of Science:
- Medical Imaging
- Artificial Intelligence
- Infectious Diseases
Background:
- The COVID-19 pandemic necessitates rapid and accurate diagnostic tools.
- Traditional diagnostic kits have limitations, requiring supplementary methods.
- Deep learning models show promise in analyzing CT images for COVID-19 detection.
Purpose of the Study:
- To propose a novel deep learning model for enhanced COVID-19 diagnosis from CT images.
- To optimize the model's convergence time and diagnostic performance.
- To evaluate the model's effectiveness against existing deep neural networks.
Main Methods:
- A new deep learning architecture incorporating a combined pooling and Squeeze Excitation Block (SE-block) layer.
- Utilized Batch Normalization and Mish Function for optimization.
- Evaluated the model on a dataset from two public hospitals and compared it with popular deep neural networks (DNNs).
Main Results:
- The proposed model achieved a high accuracy of 99.03%.
- Demonstrated a rapid recognition time of 0.069 ms in test mode on a GPU.
- Outperformed other popular deep neural networks in classification metrics and real-time application suitability.
Conclusions:
- The novel deep learning model is highly accurate and efficient for COVID-19 diagnosis using CT scans.
- The model's architecture, optimized with SE-block, Batch Normalization, and Mish Function, offers significant advantages.
- This approach provides a valuable auxiliary tool for rapid COVID-19 detection during the pandemic.
