In this project we suggest a method for weakly supervised segmentation based on scribbles and combine it with active learning methods.

1. Start off by updating the classes list, names and dictionary in the constants.py file.
2. Either generate a pool directory from a CT volume using generate_demo_set.py or prepare a directory of images. Each image in the pool folder should have an 'image' prefix
3. Set your environment to run two processes in parallel
4. Run the gui.py gui.py file