Adaptive Successive Variational Mode Decomposition for Single-Channel EEG Artifact Removal Guided by the Proper Partition Principle
BIO Web of Conferences · 2026 · 267 · 01004
DOI: 10.1051/bioconf/202626701004 ↗Kurz gefasst
Fachartikel in der Zeitschrift BIO Web of Conferences, online erschienen am 18.11.2026. Verfasst von Yangkai Xu und 3 weiteren Autor:innen. Die Arbeit nennt 11 Literaturangaben. Der Volltext ist frei zugänglich (Open Access, CC BY 4.0).
Abstract (Originalsprache)
Electroencephalogram (EEG) signals are vital for neural analysis but easily contaminated by various artifacts like Electrooculogram (EOG). While Variational Mode Decomposition (VMD) is a powerful tool for signal denoising, its performance is primarily limited by the 2-D hyperparameter tuning for number of modes ( K ) and penalty factor (α). Successive variational mode decomposition (SVMD) solved the problem by adaptively determining parameter K , reducing the 2-D parameter searching problem to 1-D. However, applying standard SVMD can lead to suboptimal modal overlapping, compromising brainwave fidelity. In this paper, we propose an adaptive successive variational mode decomposition method guided by “proper partition” for artifacts removal from single-channel EEG. Rather than overemphasizing lossless EOG elimination, our method strategically isolates EOG energy within a single IMF while maximizing the energy preservation of the remaining clean IMFs. Then, we automatically optimize α through dynamic penalty factor search by evaluating the spectrum and calculating the correlation coefficient. Validation on the EEGDenoiseNet dataset demonstrates that the proposed adaptive SVMD significantly outperforms traditional algorithms, including WT+ICA and EEMD+PCA, in metrics such as the correlation coefficient, residual energy, and signal-to-noise ratio. This method offers a reliable approach for artifact removal, providing more reliable input signals for downstream processing pipelines.
Zitieren
APA
Xu, Y., Yang, H., Shen, Z., & Zhang, B. (2026). Adaptive Successive Variational Mode Decomposition for Single-Channel EEG Artifact Removal Guided by the Proper Partition Principle. BIO Web of Conferences, 267, 01004. https://doi.org/10.1051/bioconf/202626701004
Vancouver
Xu Y, Yang H, Shen Z, Zhang B. Adaptive Successive Variational Mode Decomposition for Single-Channel EEG Artifact Removal Guided by the Proper Partition Principle. BIO Web of Conferences. 2026;267:01004. doi:10.1051/bioconf/202626701004
BibTeX
@article{Xu2026,
author = {Xu, Yangkai and Yang, Haohan and Shen, Zhihan and Zhang, Boyang},
title = {Adaptive Successive Variational Mode Decomposition for Single-Channel EEG Artifact Removal Guided by the Proper Partition Principle},
journal = {BIO Web of Conferences},
year = {2026},
volume = {267},
pages = {01004},
doi = {10.1051/bioconf/202626701004}
}
EUVIMED