MiroSCOPE is an AI-driven platform for functional tissue unit (FTU) annotation, built as an extension to QuPath, a popular open-source software for digital pathology image analysis.
MiroSCOPE enhances QuPath by integrating advanced AI capabilities to support automated and semi-automated annotation of functional tissue units. This extension is designed to accelerate histological analysis and improve reproducibility in tissue-based studies.
- Seamless integration with QuPath
- AI-assisted FTU detection and annotation
- Support for large-scale histopathology image analysis
To use MiroSCOPE, download the appropriate build for your system from the dists folder.
You may build from the source code in this repo by following the instruction in QuPath_README.
To use MiroSCOPE, click the "MiroSCOPE" menu in the main menu bar. Choose the folder containing your images. The selected folder must contain two sub folders: images and annotations (optional). All your image files should be placed into the images folder. The annotations folder, if present, should contain annotation files for individual images. This folder structure must be strictly followed.
To test MiroSCOPE's features, we have provided a sample image set in the test-images folder:
- The annotated_image folder as an example of a previously annotated image
- The unannotated_image folder as an example of an image that has not been annotated
The source code is available in the the qupath-extension-cedar folder.
The Python-based backend used to support inference and model fine-tuning is hosted at monailabel_cedar_app. Follow the instructions there to install the backend.
If you use MiroSCOPE in your work, please cite the associated publication (TBD).
As specified in this repo.
For questions, feedback, or bug reports, please contact the development team or submit an issue through the appropriate repository.