Can I automatically filter out background or slide artifacts?

Yes, Histolab includes TissueMask and BiggestTissueBoxMask classes that automatically identify tissue regions and exclude non-informative backgrounds or pen marks.

What tile extraction strategies are available in this skill?

You can use RandomTiler for diverse sampling, GridTiler for comprehensive coverage, or ScoreTiler to rank and extract tiles based on metrics like nuclei density.

Which whole slide image formats does Histolab support?

Histolab supports standard digital pathology formats including SVS, TIFF, and NDPI, utilizing pyramidal image structures for efficient high-resolution processing.

Does Histolab handle image preprocessing and filtering?

Yes, it provides a comprehensive suite of filters for color space conversions (RGB to HED/HSV), Otsu thresholding, and morphological operations to clean images before analysis.

How can I verify the extraction locations before running a job?

The skill provides locate_mask() and locate_tiles() methods to generate visual previews of masks and tile coordinates on slide thumbnails.

Histolab Digital Pathology

Name: Histolab Digital Pathology
Author: Microck

byMicrock

•

Ciencia de Datos y ML

Automates tissue detection and tile extraction from gigapixel histopathology images for computational pathology deep learning pipelines.

Acerca de

Histolab is a specialized Claude Code skill designed to streamline digital pathology workflows by providing a robust framework for processing Whole Slide Images (WSI). It enables users to handle massive gigapixel files in formats like SVS, TIFF, and NDPI, automating the complex tasks of tissue segmentation, artifact removal, and informative tile extraction. Whether you need systematic grid coverage or quality-driven selection based on nuclei density, Histolab provides the tools to transform raw medical slides into structured datasets ready for advanced machine learning analysis.

Características Principales

81 GitHub stars
Multiple tile extraction strategies including Random, Grid, and Score-based sampling
Support for multi-format Whole Slide Images (WSI) such as SVS, TIFF, and NDPI
Extensive preprocessing filters for color space conversion and morphological operations
Automated tissue detection and mask generation to exclude background and artifacts
Integrated visualization tools for slide thumbnails, tissue masks, and tile locations

Casos de Uso

Automating the segmentation of tissue regions in gigapixel surgical pathology slides
Preparing high-quality image datasets for training histopathology deep learning models
Identifying and extracting the most informative regions of a slide using cellularity scoring

Acerca de

Características Principales

81 GitHub stars
Multiple tile extraction strategies including Random, Grid, and Score-based sampling
Support for multi-format Whole Slide Images (WSI) such as SVS, TIFF, and NDPI
Extensive preprocessing filters for color space conversion and morphological operations
Automated tissue detection and mask generation to exclude background and artifacts
Integrated visualization tools for slide thumbnails, tissue masks, and tile locations

Casos de Uso

Automating the segmentation of tissue regions in gigapixel surgical pathology slides
Preparing high-quality image datasets for training histopathology deep learning models
Identifying and extracting the most informative regions of a slide using cellularity scoring