This repository contains code that was used to reproduce and extend upon the paper "TrDosePred: A deep learning dose prediction algorithm based on transformers for head and neck cancer radiotherapy". This paper illustrates the application of 3D Vision Transformers in the field of radiation dose treatment planning. In this repository a training pipeline was set up that contains augmentation and preprocessing and allows for plug and play usage of models. For background information and detailed explanation of relevant topics take a look at the blogpost.
This code was developed using Python 3.11. To setup the project, clone the repository and install the requirements.
git clone [email protected]:oxkitsune/DL2.git
pip install -r requirements.txt
The code can be run using the following command:
python -m src.main
To use see all available options use the --help flag.
This section contains commands that were used to achieve resultst presented in the blogpost.
Experiments for the three-layer UNETR architecture:
python -m src.main --batch-size 4 --loss mae --parallel --model unetr --lr 0.0004
python -m src.main --batch-size 4 --loss dvh --parallel --model unetr --lr 0.0004
python -m src.main --batch-size 4 --loss moment --parallel --model unetr --lr 0.0004
python -m src.main --batch-size 4 --loss all --parallel --model unetr --lr 0.0004
Experiments for the four-layer UNETR architecture:
python -m src.main --batch-size 4 --loss mae --parallel --model bigunetr --lr 0.0004
Experiments for the RNN-based architecture:
python -m src.main --batch-size 4 --loss mae --parallel --model rnnunetr --lr 0.0004