Related Experiment Video
Updated: Jul 8, 2025
02:09
Multi-modal Pulmonary Imaging: Using Complementary Information from CT and Hyperpolarized 129Xe MRI to Evaluate Lung Structure-Function
Published on: April 12, 2024
616
SARS-CoV-2 Detection: Radiology based Multi-modal Multi-task Framework
Summary
This study introduces a novel deep learning framework for detecting SARS-CoV-2 using chest X-rays or CT scans. The multi-modal approach achieves high accuracy in identifying COVID-19 patients.
Area of Science:
- Medical Imaging
- Artificial Intelligence
- Infectious Diseases
Background:
- The COVID-19 pandemic necessitates rapid and accurate diagnostic tools.
- Radiology, including X-rays and CT scans, plays a crucial role in early SARS-CoV-2 detection.
- Existing deep learning models often focus on a single imaging modality, limiting their comprehensive application.
Purpose of the Study:
- To develop and evaluate a multi-modal, multi-task learning framework for SARS-CoV-2 detection.
- To leverage both X-ray and CT scan data for improved diagnostic accuracy.
- To create a flexible framework adaptable to different imaging inputs.
Main Methods:
- A multi-modal, multi-task learning framework was designed.
- The framework utilizes a shared feature embedding for common information across imaging types.
- Task-specific embeddings were developed for modality-independent features, combined for classification.
Main Results:
- The framework achieved high accuracy in detecting SARS-CoV-2.
- Accuracy rates of 98.23% using X-rays and 98.83% using CT scans were reported.
- The multi-modal approach demonstrated robust performance across different imaging modalities.
Conclusions:
- The proposed framework effectively identifies SARS-CoV-2 patients using either X-rays or CT scans.
- Multi-modal learning enhances diagnostic performance in COVID-19 detection.
- This approach offers a promising tool for radiological screening of SARS-CoV-2.
