Pulse2Pulse (DeepFake ECG GAN)

SimulaMet / Oslo Metropolitan University (Thambawita et al.)

12-lead ECG

Filter catalog by Modality:
ECG

General Purpose / Multi-task

Filter catalog by Disease / Trait:
General / Foundation

Generation

Filter catalog by Task Type:
Generation

CNN (1D)

Filter catalog by Architecture:
Convolutional (CNN)

PyTorch

Filter catalog by Framework:
PyTorch

MIT

Filter catalog by License:
Permissive

file_copy

View model files code View code

Generative adversarial network that synthesizes realistic 10-second, 12-lead normal-sinus-rhythm ECGs from scratch, without using any real patient data at inference time, enabling privacy-preserving data sharing and augmentation. Uses a U-Net-style 1D deconvolutional generator with a WaveGAN-inspired discriminator. Outperformed a WaveGAN* baseline on the fraction of generated tracings classified as normal sinus rhythm by a commercial ECG interpretation algorithm. Developed by SimulaMet and Oslo Metropolitan University.

memory Specifications

gavel License

MIT

check_small Open source check_small Commercial use OK check_small Modifications allowed Attribution required No share-alike requirement

License for model weights only. Associated code may be licensed seperately, check code source for specific terms.

description Publication

DeepFake electrocardiograms using generative adversarial networks are the beginning of the end for privacy issues in medicine open_in_new

2021 · original paper

DOI: 10.1038/s41598-021-01295-2

database Training & evaluation data

GESUS + Inter99 Population Cohort ECGs

train

public 7,233 subjects · Denmark

7,233 real normal-sinus-rhythm 12-lead ECGs drawn from the Danish GESUS and Inter99 population cohorts, used as real training data for the generator.

science Capabilities & performance

Synthetic 10-second 12-lead normal-sinus-rhythm ECG waveform generated de novo

Generation General Purpose / Multi-task

CVAI Catalog

View Catalog