Code for the paper "Segmentation of Roots in Soil with U-Net"

Data is available from https://zenodo.org/record/3527713

Trained model is available from https://zenodo.org/record/3484015

Install dependencies

> pip install -r requirements.txt

Compute and print U-Net test set metrics

> cd ./src
> python ./unet/test.py

Compute Frangi test set metrics

> cd ./src
> python ./frangi/test.py

Note:

I'm currently recommending the newer RootPainter software for new segmentation projects:

Paper available on bioarxiv: https://www.biorxiv.org/content/10.1101/2020.04.16.044461v2

Paper avilable on researchgate: https://www.researchgate.net/publication/340765796_RootPainter_Deep_Learning_Segmentation_of_Biological_Images_with_Corrective_Annotation

Code and software available on github: https://github.com/Abe404/root_painter