histolab — обработка патологических слайдов WSI

★ 7.0 · media-processing

histolab is a Claude Code skill that automates whole slide image (WSI) processing in Python, covering tissue detection, tile extraction from gigapixel slides, and dataset preparation for deep learning in digital pathology. Three tiling strategies are provided — RandomTiler for sampling, GridTiler for full coverage, and ScoreTiler for quality-driven selection — alongside Reinhard and Macenko stain normalization for H&E images. Six documented capability areas include slide management, tissue masks (TissueMask and BiggestTissueBoxMask), morphological filters, preprocessing, visualization, and five end-to-end workflows. Compatible with Python 3.8–3.11 on Linux and macOS, it requires OpenSlide system libraries and optionally pooch for built-in TCGA sample slides. Best suited for straightforward histopathology pipelines, dataset curation, and tile-based analysis without advanced multiplexed or spatial proteomics imaging.