Deep learning based bio-metric authentication system using a high temporal/frequency resolution transform
2024

Deep Learning for ECG-Based Identity Verification

Sample size: 66 publication 10 minutes Evidence: high

Author Information

Author(s): Maleki Lonbar Sajjad, Beigi Akram, Bagheri Nasour, Peris-Lopez Pedro, Camara Carmen

Primary Institution: Shahid Rajaee Teacher Training University

Hypothesis

Can ECG signals be effectively used for identity verification through deep learning techniques?

Conclusion

The study demonstrates that ECG signals can achieve high accuracy in identity verification using deep learning methods.

Supporting Evidence

  • The model achieved an accuracy of 99.3% on the NSRDB dataset.
  • The model demonstrated an accuracy of 99.004% on the MITDB dataset.
  • The Equal Error Rate (EER) was 0.8% for both datasets.
  • Deep learning techniques significantly improved the performance of ECG-based authentication.

Takeaway

This study shows that your heart's electrical signals can be used to confirm your identity, just like a fingerprint.

Methodology

The study used ECG signals processed through the Wigner-Ville distribution and analyzed with a convolutional neural network (GoogleNet) for identity verification.

Limitations

The study may be limited by the noise present in ECG signals and the need for extensive preprocessing.

Participant Demographics

The study included 48 individuals from the MITDB dataset and 18 from the NSRDB dataset, with ages ranging from 20 to 89 years.

Statistical Information

P-Value

0.8

Statistical Significance

p<0.05

Digital Object Identifier (DOI)

10.3389/fdgth.2024.1463713

Want to read the original?

Access the complete publication on the publisher's website

View Original Publication