Time-Varying Graph Signal Reconstruction With Temporal Filter Identification
IEEE Access · 2026 · 14 · 70265-70278
DOI: 10.1109/access.2026.3690332 ↗Kurz gefasst
Fachartikel in der Zeitschrift IEEE Access, 2026. Verfasst von Shuyang Feng und 2 weiteren Autor:innen. Die Arbeit nennt 45 Literaturangaben. Der Volltext ist frei zugänglich (Open Access, CC BY 4.0).
Abstract (Originalsprache)
Graph signal reconstruction aims to recover signals from incomplete or noisy observations by exploiting correlations induced by the underlying graph topology. Although recent works have extended reconstruction methods from static or band-limited graph signals to time-varying settings, many existing approaches rely on a priori smoothness assumptions on temporal signal differences, which correspond to restrictive temporal filter structures. To enhance flexibility and applicability, a temporal filtering framework that does not rely on a predefined temporal filtering model is proposed for the joint reconstruction of time-varying graph signals and the estimation of temporal filters. To address the practical challenge of unknown filter order, a joint signal reconstruction and filter identification method is further developed to simultaneously estimate both filter coefficients and their effective order. The resulting optimization problems are efficiently solved using the alternating direction method of multipliers (ADMM). Experimental results demonstrate the effectiveness of the proposed approach in capturing both nodal characteristics and the temporal evolution of graph signals.
Zitieren
APA
Feng, S., Yang, R., & Shi, X. (2026). Time-Varying Graph Signal Reconstruction With Temporal Filter Identification. IEEE Access, 14, 70265-70278. https://doi.org/10.1109/access.2026.3690332
Vancouver
Feng S, Yang R, Shi X. Time-Varying Graph Signal Reconstruction With Temporal Filter Identification. IEEE Access. 2026;14:70265-70278. doi:10.1109/access.2026.3690332
BibTeX
@article{Feng2026,
author = {Feng, Shuyang and Yang, Rui and Shi, Xintong},
title = {Time-Varying Graph Signal Reconstruction With Temporal Filter Identification},
journal = {IEEE Access},
year = {2026},
volume = {14},
pages = {70265--70278},
doi = {10.1109/access.2026.3690332}
}
EUVIMED