Multiscale tools library

A library that includes four components: preprocessing tools, regressor tools, classification tools, and segmentation tools.

  • The preprocessing tools include scripts to extract the patches from several WSI’s scales.
  • The scale regressor tools include scripts to generate the input data, train, test and use a scale regressor.
  • The classification tools include scripts for training a CNN in a fully supervised fashion.
  • The segmentation tools include models such as HookNet for the semantic segmentation in histopathology whole-slide images.

The repository is referencable through the paper: Marini, N., Otálora, S., Podareanu, D., van Rijthoven, M., van der Laak, J., Ciompi, F., Müller H. & Atzori, M. (2021). Multi_Scale_Tools: a python library to exploit multi-scale whole slide images. Frontiers in Computer Science, 3, 68.

The library is available on its GitHub repository.