COVID-19 Detection in CT/X-ray Imagery Using Vision Transformers

Area of Science:

• Medical Imaging
• Artificial Intelligence
• Deep Learning

Background:

• The 2019 Coronavirus disease (COVID-19) pandemic has caused significant global health and economic disruption.
• Medical imaging, including Computed Tomography (CT) and X-ray, plays a crucial role in disease screening.
• Deep learning models offer advanced capabilities for analyzing medical images.

Purpose of the Study:

• To propose a novel deep learning framework for enhanced Coronavirus detection using CT and X-ray images.
• To leverage a Vision Transformer architecture with a Siamese encoder for improved image analysis.

Main Methods:

• A Vision Transformer architecture with a Siamese encoder was employed.
• The encoder utilized two branches to process original and augmented image views.
• Input images were divided into patches and processed through the encoder.

Main Results:

• The proposed framework achieved superior performance over state-of-the-art methods on public CT and X-ray datasets.
• Key performance metrics included accuracy, precision, recall, specificity, and F1 score.
• The system demonstrated robustness even with limited training data.

Conclusions:

• The developed deep learning framework shows significant potential for accurate and reliable Coronavirus detection.
• The approach offers a robust solution for medical image analysis in pandemic scenarios.
• This method advances the application of AI in diagnostic imaging for infectious diseases.