evaluation

Evaluation script

Download data

In order to reproduce the results reported in End-To-End Learned Random Walker for Seeded Image Segmentation please download cvpr_2109_lrw.h5 and baseline_lrw.h5 from: https://heibox.uni-heidelberg.de/published/cvpr2019_lrw/

installation

conda env create -f ./environment.yml 
conda activate py36_cremi
pip install git+https://github.com/lorenzocerrone/cremi_python.git

Run Learned RW evaluation

python evaluation_test_cremi.py --gtpath ./cvpr_2109_lrw.h5 --segpath ./cvpr_2109_lrw.h5 

Run baseline evaluation

standard RandomWalker algorithm

python evaluation_test_cremi.py --gtpath ./cvpr_2109_lrw.h5 --segpath ./baseline_lrw.h5 --segdataset segmentation_stRW

standard Watershed algorithm

python evaluation_test_cremi.py --gtpath ./cvpr_2109_lrw.h5 --segpath ./baseline_lrw.h5 --segdataset segmentation_stWS