COVID-CT-MD, COVID-19 computed tomography scan dataset applicable in machine learning and deep learning

Area of Science:

• Medical Imaging
• Artificial Intelligence in Medicine
• Infectious Diseases

Background:

• The COVID-19 pandemic has severely impacted global health systems.
• Current diagnostic methods like RT-PCR have limitations in speed and sensitivity.
• Computed Tomography (CT) offers higher sensitivity for diagnosing COVID-19 compared to Chest Radiography (CXR).

Purpose of the Study:

• To introduce the COVID-CT-MD, a novel dataset for COVID-19 research.
• To provide a comprehensive dataset including COVID-19, healthy, and Community Acquired Pneumonia (CAP) cases.
• To facilitate the development of advanced Machine Learning (ML) and Deep Neural Network (DNN) models for COVID-19 diagnosis.

Main Methods:

• Creation of the COVID-CT-MD dataset comprising CT scans.
• Inclusion of diverse patient groups: COVID-19 positive, healthy, and CAP.
• Annotation of the dataset at lobe-level, slice-level, and patient-level.

Main Results:

• The COVID-CT-MD dataset is now available for research purposes.
• The dataset includes detailed labels crucial for training ML/DNN models.
• It offers a valuable resource for comparative studies of respiratory diseases.

Conclusions:

• The COVID-CT-MD dataset is a significant contribution to COVID-19 diagnostic research.
• It has the potential to accelerate the development of AI-driven diagnostic tools.
• This dataset can aid in differentiating COVID-19 from other pulmonary conditions like CAP.