Can I use this for CODEX imaging analysis?

Yes, PathML provides specialized slide classes and transforms for processing multiparametric data from platforms like CODEX, Vectra, and MERFISH.

Does this skill help with image stain normalization?

Yes, it includes automated pipelines for H&E stain normalization using established methods like Macenko and Vahadane to ensure consistency across datasets.

Which slide formats does PathML support?

PathML supports over 160 proprietary and open formats, including common types like Aperio SVS, Hamamatsu NDPI, Leica SCN, DICOM, and OME-TIFF.

Can I train deep learning models directly with this skill?

Absolutely. PathML integrates with PyTorch and provides pre-built architectures like HoVer-Net for simultaneous nucleus segmentation and classification.

PathML Computational Pathology

Name: PathML Computational Pathology
Author: ovachiever

byovachiever

•

データサイエンスとML

Automates computational pathology workflows by processing whole-slide images and multiparametric data for machine learning analysis.

PathML is a modular Python toolkit designed for the deep analysis of pathology imaging data, enabling researchers to load, preprocess, and analyze whole-slide images (WSI) at scale. It bridges the gap between raw medical imaging and advanced machine learning by providing specialized tools for nucleus detection, spatial graph construction, and stain normalization across 160+ proprietary formats. Whether working with standard H&E slides or complex multiplex immunofluorescence like CODEX, this skill streamlines the entire research pipeline from raw image ingestion to deploying models like HoVer-Net for tissue segmentation and spatial proteomics.

主な機能

01Support for 160+ proprietary slide formats including SVS, NDPI, DICOM, and OME-TIFF

02Specialized modules for multiparametric imaging analysis and CODEX workflows

03Built-in deep learning support for HoVer-Net and HACTNet model architectures

04Automated preprocessing pipelines featuring Macenko/Vahadane stain normalization

0515 GitHub stars

06Spatial graph construction for representing cellular and tissue-level relationships

ユースケース

01Preprocessing and normalizing H&E slides for large-scale clinical research datasets

02Conducting spatial proteomics analysis on multiplexed CODEX and Vectra imaging

03Automating nucleus detection, segmentation, and classification for digital pathology

What are Skills?·How to Install

Install with 🐟 Skill.Fish

npx skillfish add ovachiever/droid-tings pathml

For use in Claude.ai and ChatGPT

Download Skill