سامانه همایش های دانشگاه آزاد اسلامی واحد نجف آباد

صفحه اصلی / اولین همایش بین المللی تحول خدمات سلامت: تلفیق مراقبت، درمان، بازتوانی و امیدبخشی

Novel Artifact Removal and Noise Reduction Techniques for Vital Signals in Clinical Environments: A Systematic Review

نویسندگان :

Bita Babaei¹ Alireza Zarei²

1- موسسه آموزش عالی آپادانا 2- دانشگاه بین المللی امام رضا علیه السلام

کلمات کلیدی :

Electrocardiography،Artifacts،Signal-To-Noise Ratio،Artificial Intelligence،Signal Processing،Computer-Assisted،Telemedicine

چکیده :

Abstract: Background: Wearable electrocardiogram (ECG) devices are pivotal for the continuous monitoring and management of cardiovascular diseases. However, motion artifacts and environmental noise severely compromise signal integrity in real-world settings, necessitating robust noise reduction techniques to ensure clinical reliability and diagnostic accuracy. Therefore, the primary objective of this systematic review is to evaluate and compare traditional and artificial intelligence (AI)-based artifact reduction techniques for wearable ECG signals, and to identify their limitations and potential for real-world clinical applications. Method: A systematic review was conducted following PRISMA guidelines. Major databases (PubMed, IEEE Xplore, Web of Science, Scopus) were comprehensively searched for studies published between 2012 and 2025 focusing on ECG denoising. Based on rigorous inclusion criteria, 11 high-quality studies were strategically selected to evaluate both traditional and artificial intelligence (AI)-based signal processing methods across time, frequency, and time-frequency domains. Finding: Traditional techniques, such as adaptive filtering and wavelet transforms, effectively mitigate baseline wander but often struggle with complex motion artifacts that overlap with the ECG frequency spectrum. Conversely, the integration of AI-particularly deep learning architectures like autoencoders and convolutional neural networks-demonstrated superior performance in non-linear noise extraction and signal reconstruction. Despite their immense potential, AI models currently face challenges regarding computational overhead on low-power wearables and clinical interpretability. Hybrid frameworks showed the most promise in balancing accuracy and computational efficiency. Conclusion: Effective artifact reduction is crucial for the sustainable deployment of remote cardiac care. Integrating lightweight, edge-AI algorithms with traditional processing strategies offers a viable pathway for high-fidelity wearable ECG monitoring. Overcoming current limitations requires standardized noisy datasets, optimized algorithmic frameworks, and enhanced hardware capabilities.