IMFSegNet: Cost-effective and objective quantification of intramuscular fat in histological sections by deep learning

Jan-Philipp Praetorius, Kassandra Walluks, Carl-Magnus Svensson, Dirk Arnold, Marc Thilo Figge

Submitted

A comparative assessment to quantify intramuscular fat in H&E stained tissue sections

IMFSegNet

Please visit github repository for more information.

Contents

  • data contains all raw images, predictions and quantifications derived from the images
  • examples contains example python scripts for the cross-validation and the corresponding training-test data
  • model contains all embedded image analysis workflows with the corresponding versions for JIPipe, ilastik, kmeans clustering and the IMFSegNet
  • sheepfat contains all python scripts used to build and train the SegNet architecture as well as predict with it and the kmeans clustering
  • SheepFat.jip contains the JIPipe project file which is used to process the raw images into quantitative results