12,058 chest CT scan images from 638 patients categorized into COVID-19 and normal cases. The dataset includes 2,282 images from 95 COVID-19 patients and 6,396 images from 282 normal patients, all sourced from the National Research Institute of Tuberculosis and Lung Disease. These images are provided in high-resolution 16-bit format to maintain clinical accuracy.
Use Cases
- Train a binary classifier to detect COVID-19 using the patient-level labels and CT image slices
- Benchmark the performance of deep learning architectures on high-bit-depth medical imaging data
- Analyze radiological features of COVID-19 by comparing the 2,282 positive images against the 6,396 normal images
Strengths
- 12,058 total CT scan images from 638 unique patients
- 2,282 images from 95 COVID-19 patients and 6,396 images from 282 normal patients
- Images are provided in 16-bit TIFF format
- Sourced from the National Research Institute of Tuberculosis and Lung Disease (NRITLD)