Deepfake_ecg Open-source Model - Free Generation of Realistic ECG Data to Boost Medical Research and Development

Deepfake Ecg

Developed by deepsynthbody

A deep learning model for generating synthetic ECG that creates realistic electrocardiogram data for medical research and development

Molecular Model

Transformers

English#Synthetic ECG generation #Medical data augmentation #Unconditional image generation

Downloads 21

Release Time : 2/24/2023

Model Overview

This model generates high-quality synthetic ECG data through deep learning technology, applicable for medical AI training, privacy protection research, and other scenarios

Model Features

High-quality synthetic ECG

Capable of generating realistic ECG data that can replace real patient data

Large-scale data generation

Pre-generated dataset of 150,000 ECG records with support for further expansion

Privacy protection applications

Synthetic data can be used for medical AI development without involving real patient privacy

Model Capabilities

ECG generation

Medical data synthesis

Unconditional image generation

Use Cases

Medical research

AI ECG analysis model training

Using synthetic ECG data to train medical AI models

Avoids privacy issues associated with using real patient data

Medical algorithm development

Providing large amounts of test data for ECG analysis algorithms

Accelerates development cycles and reduces data acquisition costs

Privacy protection

Anonymized medical research

Using synthetic data instead of real patient data for research

Completely eliminates the risk of patient privacy leaks

🚀 Deepfake-ECG

This project focuses on generating synthetic ECGs using deep learning techniques, offering solutions for privacy - related issues in medical data.

🚀 Quick Start

Generate Synthetic ECGs from Hugging Face

from transformers import AutoModel

model = AutoModel.from_pretrained("deepsynthbody/deepfake_ecg", trust_remote_code=True)

out = model(num_samples=5)

Train the Model from Scratch

If you want to train the model from scratch, please refer to our development repository Pulse2Pulse.

✨ Features

Generate DeepFake ECGs with 8 - lead values for 10s (5000 values per lead).
Convert 8 - leads format to 12 - leads format using specific equations.
Provide pre - generated DeepFake ECGs and corresponding MUSE reports.

💻 Usage Examples

Basic Usage

The generator functions can generate DeepFake ECGs with 8 - lead values [lead names from first column to eighth column: 'I','II','V1','V2','V3','V4','V5','V6'] for 10s (5000 values per lead). These 8 - leads format can be converted to 12 - leads format using the following equations:

lead III value = (lead II value) - (lead I value)
lead aVR value = -0.5*(lead I value + lead II value)
lead aVL value = lead I value - 0.5 * lead II value
lead aVF value = lead II value - 0.5 * lead I value

Pre - generated Datasets

Pre - generated DeepFake ECGs and corresponding MUSE reports are available at https://osf.io/6hved/ or https://huggingface.co/datasets/deepsynthbody/deepfake_ecg.

In this repository, there are two DeepFake datasets:
1. 150k dataset - Randomly generated 150k DeepFakeECGs
2. Filtered all normals dataset - Only "Normal" ECGs filtered using the MUSE analysis report

Visual Examples

A real ECG vs a DeepFake ECG (from left to right):

Real vs Fake

A sample DeepFake ECG:

A regenerated sample

📚 Documentation

🤝 Contributing

Pull requests are welcome. For major changes, please open an issue first to discuss what you would like to change. Please make sure to update tests as appropriate.

📄 Citation

@article{thambawita2021deepfake,
  title={DeepFake electrocardiograms using generative adversarial networks are the beginning of the end for privacy issues in medicine},
  author={Thambawita, Vajira and Isaksen, Jonas L and Hicks, Steven A and Ghouse, Jonas and Ahlberg, Gustav and Linneberg, Allan and Grarup, Niels and Ellervik, Christina and Olesen, Morten Salling and Hansen, Torben and others},
  journal={Scientific reports},
  volume={11},
  number={1},
  pages={1--8},
  year={2021},
  publisher={Nature Publishing Group}
}

📄 License

MIT

📞 Contact

For more details, please contact: vajira@simula.no, michael@simula.no

Featured Recommended AI Models

Empowering the Future, Your AI Solution Knowledge Base

English 简体中文繁體中文にほんご