200,000 labeled nuclei instances across 19 distinct tissue types facilitate instance segmentation and classification tasks. This dataset represents the first part of a three-part series focused on histopathology research.
Use Cases
- Train instance segmentation models to detect individual nuclei boundaries using the provided labels
- Develop classification algorithms to categorize nuclei based on the 19 tissue type labels
- Evaluate model performance on multi-organ cancer detection using the diverse tissue categories
Strengths
- 200,000 labeled nuclei instances for segmentation and classification
- Includes 19 distinct tissue types for cross-organ generalization
- First installment of a three-part series providing large-scale histopathology data