This repository provides the functions and notebooks which were used to perform test and evaluation for the paper "Independent assessment of a deep learning system for lymph node metastasis detection on the Augmented Reality Microscope". Note that the code provided here is only for reference; it will not work without the corresponding data and ground truth files.

The data folder contains inference results from the entire test set in .csv format at 10x, 20x, and 40x resolution. For each annotated region of interest (ROI), the csv includes its location on its field of view and whole slide image, ground truth label, subclass, and model inference. The data files are provided in zipped format due to the large file size.