Detection of Intelligent Tampering in Wireless Electrocardiogram Signals Using Hybrid Machine Learning

2025-07-08

AI Summary

To address the vulnerability of wireless electrocardiogram (ECG) signals to intelligent tampering in health monitoring and biometric authentication, this paper proposes a tampering detection and identity verification framework based on time-frequency representation and hybrid deep learning. First, continuous wavelet transform (CWT) converts 1D ECG signals into 2D time-frequency images. Second, a FeatCNN-TranCNN classification model—integrating convolutional neural networks (CNNs) and Transformers—is developed for tampering detection; additionally, a CNN-Transformer Siamese network enables fine-grained identity verification. Third, diverse simulated attack strategies are introduced to emulate realistic tampering scenarios. Experiments demonstrate detection accuracy exceeding 99.5% under highly fragmented tampering, an average accuracy of 98% for subtle tampering, and 100% identity verification accuracy. The core contribution lies in the synergistic optimization of time-frequency modeling and multimodal architecture, significantly enhancing robustness and generalization capability.

Technology Category

• Computer Vision: Multi-modal Vision
• Machine Learning: Calibration & Uncertainty Quantification
• Planning, Routing, and Scheduling: Activity and Plan Recognition

Application Category

• User Modeling, Personalization and Recommendation: Attacks and countermeasures in recommendation systems
• Security and Privacy: Data transparency and provenance
• Systems and Infrastructure for Web, Mobile and WoT: Sustainability and carbon-aware systems for Web, mobile, and WoT

Abstract

With the proliferation of wireless electrocardiogram (ECG) systems for health monitoring and authentication, protecting signal integrity against tampering is becoming increasingly important. This paper analyzes the performance of CNN, ResNet, and hybrid Transformer-CNN models for tamper detection. It also evaluates the performance of a Siamese network for ECG based identity verification. Six tampering strategies, including structured segment substitutions and random insertions, are emulated to mimic real world attacks. The one-dimensional ECG signals are transformed into a two dimensional representation in the time frequency domain using the continuous wavelet transform (CWT). The models are trained and evaluated using ECG data from 54 subjects recorded in four sessions 2019 to 2025 outside of clinical settings while the subjects performed seven different daily activities. Experimental results show that in highly fragmented manipulation scenarios, CNN, FeatCNN-TranCNN, FeatCNN-Tran and ResNet models achieved an accuracy exceeding 99.5 percent . Similarly, for subtle manipulations (for example, 50 percent from A and 50 percent from B and, 75 percent from A and 25 percent from B substitutions) our FeatCNN-TranCNN model demonstrated consistently reliable performance, achieving an average accuracy of 98 percent . For identity verification, the pure Transformer-Siamese network achieved an average accuracy of 98.30 percent . In contrast, the hybrid CNN-Transformer Siamese model delivered perfect verification performance with 100 percent accuracy.

Problem

• Detect intelligent tampering in wireless ECG signals
• Evaluate hybrid ML models for ECG tamper detection
• Assess ECG identity verification using Siamese networks

Innovation

• Hybrid Transformer-CNN models for tamper detection
• Siamese network for ECG identity verification
• Continuous wavelet transform for signal representation