Ahcore are the AI for Oncology core components for computational pathology. It provides a set of tools for working with pathology images and annotations. It also offers standard computational pathology algorithms.
Check the full documentation for more details on how to use ahcore.
Ahcore is not intended for clinical use. It is licensed under the Apache License 2.0.
Ahcore comes with several models included, check the model zoo for more information:
- Tissue/background segmentation in H&Es
To train a model, first make sure ahcore is installed:
git clone https://github.com/NKI-AI/ahcore.git && cd ahcore && pip install -e .
Next, make sure that the environmental variables are set correctly. We are using dotenv for this
but you can set the variables themselves. Check tools/.env.example to see which variables you need to set.
If you use dotenv, you will need to make a copy cp tools/.env.example tools/.env and fill in the correct values.
Next, you will need to create manifest, and once done, it's often beneficial to copy the data to a local drive, if its
not there yet. You can use the ahcore data copy-data-from-manifest tool for this.
