This repository contains code examples for working with our publicly available dataset on microscopic whole slide images (WSIs) of canine cutaneous tumors. The dataset can be downloaded here via the website of The Cancer Imaging Archive.
We provide two Jupyter notebooks for training and applying a segmentation network to the dataset:
- Training: segmentation_training.ipynb
- Inference: segmentation_inference.ipynb
We provide two Jupyter notebooks for training and applying a classification network to the dataset:
- Training: classification_training.ipynb
- Inference: classification_inference.ipynb
We provide a Jupyter notebook for WSI inference and performance evaluation:
We provide two SlideRunner PlugIns for visualization of the segmentation and classification results.
We provide two pre-trained models for patch segmentation and classification. These can be found in the models folder.
We provide six python modules to convert various annotation formats into one another. These can be found in the annotation_conversion folder.