Mild Cognitive Impairment Detection from Rey-Osterrieth Complex Figure Copy Drawings using a Contrastive Loss Siamese Neural Network

This code repository is the official source code of the paper "Mild Cognitive Impairment Detection from Rey-Osterrieth Complex Figure Copy Drawings using a Contrastive Loss Siamese Neural Network" by Juan Guerrero Martín et al.

Requirements

Operating system: GNU/Linux Debian 13 (trixie=stable 2025-08-09)

Hardware environment: Intel(R) Core(TM) i7-7820X CPU @ 3.60Ghz, 32 GB RAM, NVIDIA GeForce GTX 1080 GPU.

Programming language: Python 3.8.12

Programming libraries: TensorFlow + Keras 2.4.1

Download the augmented ROCFD528 (binary images) dataset.

The default directory of the dataset is: /home/jguerrero/Desarrollo/DATA/proyecto_REY/datasets/rocfd528_augmented/

Please note that the files data/dataset_information/rocfd528_info.csv (ROCFD528 labels) and data/dataset_information/rocfd528_augmented_info.csv (augmented ROCFD528 labels) are only available upon formal request.

Usage

# 1. Choose your workspace and download our repository.
cd ${CUSTOMIZED_WORKSPACE}
git clone https://github.com/SIMDA-UNED/rocf-mci-detection.git

# 2. Enter the directory.
cd rocf-mci-detection

# 3. Convert our dataset into a pickle.
python utils/dataset_to_pickle.py

# 4. Execute any of our scripts.

Example:

cd training
python train_siamese_model_with_rocf_dataset.py

Citations

If you find this code useful to your research, please cite our paper as the following bibtex:

@article{guerrero2025mild,
  title={Mild Cognitive Impairment Detection from Rey-Osterrieth Complex Figure Copy Drawings Using a Contrastive Loss Siamese Neural Network},
  author={Guerrero-Mart{\'\i}n, Juan and Estella-Nonay Eladio and Bachiller-Mayoral, Margarita and Rinc{\'o}n, Mariano},
  journal={Computers, Materials \& Continua},
  volume={85},
  number={3},
  pages={4729--4752},
  year={2025},
  publisher={Tech Science Press}
}

License

This project is licensed under the GNU General Public License v3.0.

Funding

This research has been supported by an FPI-UNED-2021 scholarship.

Contact

If you would have any discussion on this code repository, please feel free to send an email to Juan Guerrero Martín.
Email: jguerrero@dia.uned.es