nicholaslaw / tumor-replicator

Replicating and Generating Another Tumor in An Image of the Brain

Installation

Open terminal in this directory and run either:

1. pip install the module itself

pip install .

2. setup.py (same as 1)

python setup.py install

3. docker-compose to set up an environment with module installed

docker-compose up -d

Getting Started

You would need two images before using this package:

1. Original Image

2. Tumor Mask

from tureplicator.replicate import TumorReplicator

r = TumorReplicator(tumor_box_size=50, skull_threshold=190, outer_erosion=50, inner_erosion=5,
                   spawn_tol=0.6, overlap_range=(0.4,0.5), dist_transform_mask_size=5, dist_transform_scalar=1.5,
                   seed=42)
result = r.generate(tumor, mask)

After running the generate() method, you should be able to derive results similar to this:

Example Notebooks

Example notebooks provided will demonstrate derivation of idea and how to use the package.

Jupyter Notebook Server

To set up a notebook server, follow step 3 of Installation and assuming default settings are applied, head to http://localhost:8889/tree to view existing or create new notebooks to perform experiments with the module. Token would be password by default.