A review of deep learning-based detection methods for COVID-19

Area of Science:

• Medical Imaging
• Artificial Intelligence
• Infectious Diseases

Background:

• The COVID-19 pandemic necessitates rapid and accurate diagnostic tools.
• Lung imaging, including Chest X-ray (CXR) and computed tomography (CT), plays a vital role in detecting coronavirus infection.
• Deep learning demonstrates significant potential in medical image analysis and computer vision tasks.

Purpose of the Study:

• To survey current deep learning methodologies for COVID-19 detection in lung images.
• To provide a comprehensive overview of available deep learning techniques, public datasets, and their applications.
• To compare evaluation metrics used in deep learning-based COVID-19 detection for future research guidance.

Main Methods:

• Literature review of deep learning methods applied to COVID-19 detection in lung images.
• Categorization of methodologies based on their approach to image analysis.
• Summary of public datasets utilized and specific datasets associated with each method.

Main Results:

• Identification and summarization of various deep learning architectures and algorithms used for COVID-19 detection.
• Compilation of commonly used public datasets for training and validation.
• Comparative analysis of evaluation metrics (e.g., accuracy, sensitivity, specificity) across different methods.

Conclusions:

• Deep learning offers promising avenues for automated COVID-19 detection from lung images.
• A consolidated understanding of methods, datasets, and metrics is crucial for advancing research and clinical application.
• This survey serves as a valuable resource for researchers developing and evaluating AI-driven diagnostic tools for COVID-19.