RadiomicsFill-Mammo: Synthetic Mammogram Mass Manipulation with Radiomics Features

🔥 Official implementation of "RadiomicsFill-Mammo: Synthetic Mammogram Mass Manipulation with Radiomics Features" (MICCAI 2024 Spotlight🎉)

Datasets

This project utilizes the VinDr-Mammo and INbreast datasets. You can download these datasets from the following links:

• VinDr-Mammo
• INbreast

Usage

Installation

To set up your development environment, follow the steps below:

1. Pull the Docker image:

   We are using the pytorch/pytorch:2.2.1-cuda11.8-cudnn8-devel Docker image. You can pull it from Docker Hub by running:

   docker pull pytorch/pytorch:2.2.1-cuda11.8-cudnn8-devel

2. Run the Docker container:

   Start a container from the pulled image. You can mount your project directory into the container for easy development:

   docker run --gpus all -it -v /path/to/your/project:/workspace --name radiomicsfill-mammo pytorch/pytorch:2.2.1-cuda11.8-cudnn8-devel /bin/bash

   Replace /path/to/your/project with the actual path to your project directory.

3. Install additional Python libraries:

   Install the required Python libraries using pip:

   pip install -r requirements.txt

Data Preprocessing

To preprocess the VinDr-Mammo dataset, run the following notebooks:

• Data Preprocessing
• Setting jsonl file

Model Training

1. Train the MET (Tabular Encoder) model:

   ./scripts/train_MET_VinDr-Mammo_embed32_enc6_dec3.sh

2. Train the RadiomicsFill-MET model:

   ./scripts/train_RadiomicsFill-MET32_VinDr-Mammo.sh

Citation

If you use this code for your research, please cite our papers.

BibTeX:

@inproceedings{na2024radiomicsfill,
  title={RadiomicsFill-Mammo: Synthetic Mammogram Mass Manipulation with Radiomics Features},
  author={Na, Inye and Kim, Jonghun and Ko, Eun Sook and Park, Hyunjin},
  booktitle={International Conference on Medical Image Computing and Computer-Assisted Intervention},
  pages={723--733},
  year={2024},
  organization={Springer}
}

Contact

For any inquiries or support, please contact us at niy0404@skku.edu.