Skip to content

Latest commit

 

History

History
7 lines (5 loc) · 624 Bytes

README.md

File metadata and controls

7 lines (5 loc) · 624 Bytes

whole_slide_images

  • big_nephro_dataset.py contains the dataset, to be used with pytorch dataloaders
  • 10Y_main.py is to be called in order to train neural networks. In particular, the MyResNet class is a neural network to evaluate multiple images together, for example several patches from a whole-slide-image.
  • 10Y_inference.py is to be called to perform inference with ensemble on an already trained network.
  • bignephroqpdata_regions.py contains several processing function used to polish the dataset and extrapolate patches from whole-slide-images. In particular, the function processing_pipeline might be useful.