Heart Signal (ECG) as Vibration Signature for Human Authentication Using TinyML

View on arXiv ← Back to list

Authors: Yogeswar Reddy Thota, Student Member, IEEE, Hanan Alsmadi, Student Member, IEEE, Tooraj Nikoubin, Senior Member, IEEE

Published: 2026-10-17 16:06:20+00:00

AI Summary

This study introduces an Electrocardiogram (ECG) based authentication framework, leveraging Tiny Machine Learning (TinyML) for deployment on resource-constrained microcontroller units (MCUs). The proposed method enables continuous, password-free authentication, addressing vulnerabilities of conventional biometrics and improving data security in the era of AI and deepfake technology. It achieves high accuracy by designing efficient TinyML models to mitigate recognition errors from noisy real-time data.

Abstract

Ensuring data security in the era of Artificial Intelligence (AI) and deepfake technology requires robust biometric systems. Conventional methods such as facial recognition, iris scans, fingerprints, and voice recognition are increasingly vulnerable to spoofing and provide only one-time authentication at login. This study introduces an Electrocardiogram (ECG) based authentication framework, compared with Photoplethysmography (PPG), leveraging Tiny Machine Learning (TinyML) for deployment on resource-constrained microcontroller units (MCUs). The proposed method enables continuous authentication, offering ongoing verification throughout device usage, which conventional biometrics cannot achieve. It also serves as a password-free solution, reducing risks of credential loss, theft, or reuse. However, ECG signals vary with physical activity, health conditions, age, and emotional states, requiring robust signal processing and learning algorithms. To address this, efficient TinyML models were designed to mitigate recognition errors from noisy real-time data. Experiments used single-lead ECG signals (Lead I) from UofTDB, MIT-BIH Arrhythmia, MIT-BIH Noise Stress Test, and PTB databases. Spectrograms and spectral features were extracted to train Deep Neural Network (DNN), 1D CNN, and 2D CNN models within the Edge Impulse platform, achieving 98% accuracy. The trained model was converted into TinyML format and deployed on an Espressif ESP-EYE (ESP32, 240 MHz). Support Vector Machine (SVM) models were also applied for benchmarking. By analyzing the number of heartbeats required for reliable recognition, the results show that averaging multiple signals improves tolerance to noise and enhances accuracy. Overall, ECG-based authentication demonstrates strong potential as a secure, real-time, and continuously verifiable method for personal identification on wearable and edge devices.

Key findings

The ECG-based authentication framework achieved up to 98% accuracy when using spectrograms and spectral features to train DNN, 1D CNN, and 2D CNN models. Averaging multiple heart signals significantly improved tolerance to noise and enhanced recognition accuracy. The DNN model utilizing spectrogram-based segment features consistently outperformed other models across different ECG databases and was effective for deployment on TinyML devices.

Approach

The proposed method preprocesses single-lead ECG signals, extracts spectrograms and spectral features, and then trains Deep Neural Network (DNN), 1D CNN, and 2D CNN models using the Edge Impulse platform. The trained models are converted into TinyML format and deployed on an Espressif ESP-EYE (ESP32) for continuous authentication, leveraging averaging of multiple signals to improve noise tolerance and accuracy.

Datasets

UofTDB, MIT-BIH Arrhythmia, MIT-BIH Noise Stress Test, PTB Diagnostic

Model(s)

Deep Neural Network (DNN), 1D CNN, 2D CNN, Support Vector Machine (SVM)

Author countries

USA

← Previous