AL_SSDA_WSI

The codes are used for the experiments of MICCAI2022
2022卒-備瀬研-荒木健吾

Directory

analysis:
To analyze about the dataset.
clustering:
Extract feature vectors from source trained model, and calculate cluster entropy.
preprocess:
To crop WSI into patches and split WSIs into train/valid/test.
S:
The codes for the experiment S that use only labeled source.
ST:
The codes for the experiment S+T that use labeled source and labeled target.
visualyze:
some functions for visualyzing featuremap and results of cluster entropy.