What problem does it solve?
This Skill automates complex computational pathology workflows, enabling advanced analysis of whole-slide images and multiparametric data without requiring deep expertise in image processing or machine learning libraries.
Core Features & Use Cases
- Image Loading: Supports over 160 WSI formats (Aperio, Hamamatsu, DICOM, OME-TIFF).
- Preprocessing: Stain normalization, tissue/nucleus detection, artifact labeling.
- Machine Learning: Integrated HoVer-Net and HACTNet for nucleus segmentation and classification.
- Spatial Analysis: Graph construction for cell-cell interactions and neighborhood analysis.
- Use Case: Analyze a batch of H&E stained cancer slides to automatically detect, segment, and classify all nuclei, then quantify tumor-infiltrating lymphocytes and their spatial distribution.
Quick Start
Use the pathml skill to load the whole-slide image at 'path/to/slide.svs' and generate tiles of size 256x256.