CrossStateECG: Multi-Scale Deep Convolutional Network with Attention for Rest-Exercise ECG Biometrics

2025-10-20

AI Summary

Existing ECG-based biometric methods achieve strong performance under resting conditions but suffer significant robustness degradation in cross-state scenarios—particularly between rest and exercise. To address this, we propose a novel ECG biometric authentication model specifically designed for rest-to-exercise and exercise-to-rest transitions. Our approach innovatively integrates multi-scale deep convolutional feature extraction with a dual-channel-and-temporal attention mechanism, enabling physiological-state-adaptive feature representation and dynamic weighting optimization. Evaluated on the Exercise-ECGID dataset, our model achieves 92.50% accuracy for rest-to-exercise identification and 94.72% for exercise-to-rest identification, with a peak accuracy of 97.85% under mixed-state conditions. Furthermore, it demonstrates strong generalizability across multiple public ECG datasets. This work constitutes the first systematic solution to the long-standing challenge of cross-physiological-state ECG biometric recognition, establishing a new paradigm for reliable identity authentication in dynamic, real-world environments.

Technology Category

• Intelligent Robots: State Estimation
• Computer Vision: Biometrics, Face, Gesture & Pose
• Machine Learning: Bio-inspired Learning

Application Category

• User Modeling, Personalization and Recommendation: Attacks and countermeasures in recommendation systems
• Security and Privacy: Large-scale security measurements
• Search and Retrieval-Augmented AI: Retrieval-Augmented Generation (RAG) and multi-modal RAG

Abstract

Current research in Electrocardiogram (ECG) biometrics mainly emphasizes resting-state conditions, leaving the performance decline in rest-exercise scenarios largely unresolved. This paper introduces CrossStateECG, a robust ECG-based authentication model explicitly tailored for cross-state (rest-exercise) conditions. The proposed model creatively combines multi-scale deep convolutional feature extraction with attention mechanisms to ensure strong identification across different physiological states. Experimental results on the exercise-ECGID dataset validate the effectiveness of CrossStateECG, achieving an identification accuracy of 92.50% in the Rest-to-Exercise scenario (training on resting ECG and testing on post-exercise ECG) and 94.72% in the Exercise-to-Rest scenario (training on post-exercise ECG and testing on resting ECG). Furthermore, CrossStateECG demonstrates exceptional performance across both state combinations, reaching an accuracy of 99.94% in Rest-to-Rest scenarios and 97.85% in Mixed-to-Mixed scenarios. Additional validations on the ECG-ID and MIT-BIH datasets further confirmed the generalization abilities of CrossStateECG, underscoring its potential as a practical solution for post-exercise ECG-based authentication in dynamic real-world settings.

Problem

• Addresses ECG biometric performance decline during rest-exercise transitions
• Develops cross-state authentication model using multi-scale convolutional networks
• Enhances identification accuracy across different physiological states

Innovation

• Multi-scale deep convolutional network for ECG feature extraction
• Attention mechanism enhances cross-state biometric identification
• Robust authentication model for rest-exercise ECG scenarios