A survey of machine learning-based methods for COVID-19 medical image analysis

Area of Science:

• Medical Imaging
• Artificial Intelligence
• Computational Biology

Background:

• The COVID-19 pandemic, caused by SARS-CoV-2, has led to millions of deaths and infections globally.
• Accurate and rapid disease detection is crucial for pandemic control.
• Lung image analysis using AI and ML has emerged as a key area for COVID-19 detection.

Purpose of the Study:

• To provide a comprehensive review of recent AI/ML-based image analysis approaches for COVID-19 detection.
• To summarize research on medical image datasets, models (ML, DL, TL), and performance metrics.
• To discuss challenges and future research directions in AI for COVID-19 analysis.

Main Methods:

• Review of recent scientific literature on COVID-19 image analysis.
• Analysis of various machine learning, deep learning, and transfer learning models.
• Examination of chest X-ray, CT, and ultrasound image datasets.

Main Results:

• Numerous AI/ML models have been developed for COVID-19 detection, classification, segmentation, and severity assessment.
• Different image modalities and datasets show varying performance metrics for these models.
• The review consolidates information on existing research, datasets, and evaluation methods.

Conclusions:

• AI and ML offer powerful tools for analyzing medical images to combat COVID-19.
• Further research is needed to address challenges and enhance the efficiency of these models.
• This review serves as a valuable resource for understanding the current landscape and future potential of AI in COVID-19 diagnostics.