194,922 CT slices from 3,745 patients categorized into normal, pneumonia, and COVID-19 classes. These images facilitate the development of deep learning models for automated infection detection within the COVID-Net open-source initiative.
Use Cases
- Train a convolutional neural network to classify infection types using the slice images and corresponding class labels
- Develop lung segmentation models to isolate regions of interest within the CT scan slices
- Evaluate model generalization across different patient demographics using the patient_id metadata
Strengths
- 194,922 CT slices across three distinct infection categories
- Data sourced from 3,745 unique patients to ensure diversity in lung pathology
- Labels include 'normal', 'pneumonia', and 'COVID-19' for multi-class classification