pathml

To build a machine learning pipeline for whole-slide image analysis, you can use this toolkit to load WSIs, preprocess data, and train models. It automates end-to-end computational pathology workflows from loading to ML training.

The best way to run graph analysis on multiparametric pathology datasets is using a modular pipeline that constructs spatial graphs. This toolkit enables tissue-level relationship mapping and spatial statistics for downstream graph neural networks.

Yes, you can process CODEX and Vectra modalities alongside H&E pathology images. The toolkit supports loading, preprocessing, and analyzing whole-slide images across these distinct imaging modalities for scalable tissue analysis.

Nucleus detection and segmentation in computational pathology works by executing end-to-end ML workflows. This toolkit handles marker quantification and supports the training, evaluation, and deployment of segmentation models directly.

You do not need external dependencies to train machine learning models for tissue segmentation, as the toolkit operates with no required dependencies. It specifies a modular pipeline architecture with optional references and assets.

Using a modular architecture for computational pathology workflows enables scalable tissue analysis and extension via a scripts directory. This approach unifies WSI loading, graph construction, and ML training into a single platform.