QuPath Annotation Exporter and ROI Splitter

This repository contains a set of scripts to export QuPath annotations as GeoJSON and split the exported ROI coordinates by cell type (classification) for further analysis.

v0.3.3

Scripts

1. ExAnnoClassifications.groovy: A Groovy script for QuPath to extract cell detection-based annotations (classifications) from QuPath TMA projects and save them as GeoJSON files.
2. qupath_roi_splitter.py: A Python script to split the ROI coordinates of QuPath TMA annotations by cell type (classification) and save them as tab-separated text files.
3. qupath_roi_splitter_folder.py: A folder version of the qupath_roi_splitter.py script that processes all GeoJSON files in a specified folder.

Requirements

• QuPath (for ExAnnoClassifications.groovy)
• Python 3.6 or later
• OpenCV (cv2)
• GeoJSON
• Pandas
• NumPy

Usage

1. Export annotations from QuPath

1. Open QuPath and load your TMA project.
2. In QuPath, go to Automate > Show script editor.
3. Copy the content of the ExAnnoClassifications.groovy script into the script editor.
4. Run the script. It will save the exported annotations as GeoJSON files in a folder named annotations_export_[TMA_name] within your QuPath project directory.

2. Split ROI coordinates by cell type (classification)

For single GeoJSON file

Run qupath_roi_splitter.py with the --qupath_roi argument pointing to the GeoJSON file you want to process. --fill is an optional argument to fill the ROI polygons. --all is an optional argument to export all features of GeoJSON (ignores if classification is present). Add --img to create a PNG for each cell type with thicker ROI outlines, a grid, and labeled X/Y axes. The PNGs are cropped to the area containing the ROIs. Axis values remain the original QuPath pixel coordinates, with Y increasing downward. PNGs from the same input share the same crop and scale.

python qupath_roi_splitter.py --qupath_roi path/to/your/geojson_file.geojson --fill

For multiple GeoJSON files in a folder

Run qupath_roi_splitter_folder.py with the qupath_roi_folder argument pointing to the folder containing the GeoJSON files you want to process. It supports the same --fill, --all, and --img options.

python qupath_roi_splitter_folder.py path/to/your/geojson_folder --fill

The scripts create a tab-separated text file for each cell type (classification) in the current working directory, named [TMA_name]_[cell_type].txt. With --img, they also create [TMA_name]_[cell_type].png plots.

License

This project is licensed under the MIT License. See the LICENSE file for details.