Robust Time-Varying Graph Signal Recovery for Dynamic Physical Sensor Network Data

About this Topic:

In this talk, the presenters will present a time-varying graph signal recovery method for estimating the true time-varying graph signal from corrupted observations by leveraging dynamic graphs. Most of the conventional methods for time-varying graph signal recovery have been proposed under the assumption that the underlying graph that houses the signals is static. However, in light of rapid advances in sensor technology, the assumption that sensor networks are time-varying like the signals is becoming a very practical problem setting. 

To this end, they focus on such cases and formulate dynamic graph signal recovery as a constrained convex optimization problem that simultaneously estimates both time-varying graph signals and sparsely modeled outliers. In our formulation, they use two types of regularizations, time-varying graph Laplacian-based and temporal difference-based, and also separately modeled missing values with known positions and unknown outliers to achieve robust estimations from highly degraded data. In addition, an algorithm is developed to efficiently solve the optimization problem based on a primal-dual splitting method. Extensive experiments on simulated drone remote sensing data and real-world sea surface temperature data demonstrate the advantages of the proposed method over existing methods.

About the Presenters:

Eisuke Yamagata received the B.E., M.E., and Ph.D. degrees in computer science from the Tokyo Institute of Technology, Tokyo, Japan, in 2020, 2022, and 2025, respectively.

He is currently an AI Research Engineer with Huawei Technologies Japan K.K., Tokyo, Japan. From 2023 to 2025, he was a Research Fellow (DC2) with the Japan Society for the Promotion of Science (JSPS). His doctoral research focused on graph signal processing. His current research interests include low-level vision, computer vision, and generative artificial intelligence.

Dr. Yamagata received the 5th IEEE SPS Tokyo Joint Chapter Student Award and the 2021 Young Encouragement Award of the Institute of Electronics, Information and Communication Engineers (IEICE) Signal Processing.

Kazuki Naganuma (M’25) received the B.E. degree from the Kanagawa Institute of Technology, Atsugi, Japan, and the M.E. and Ph.D. degrees from the Tokyo Institute of Technology, Meguro, Japan, in 2020, 2022 and 2024, respectively, all in information and computer sciences.

He is currently an Assistant Professor with the Institute of Engineering, Tokyo University of Agriculture and Technology, Fuchu, Japan. From April 2023 to March 2025, he was a Research Fellow (DC2) with the Japan Society for the Promotion of Science (JSPS), Tokyo. From October 2023 and April 2025 to present, he is a Research Fellow (PD) with JSPS and a Researcher with ACT-X, Japan Science and Technology Corporation (JST), Tokyo, Japan. His current research interests include signal and image processing and optimization theory.

Dr. Naganuma received the Student Conference Paper Award from IEEE SPS Japan Chapter in 2022, the 38th TELECOM System Technology Student Award from the Telecommunications Advancement Foundation, and the Best Paper Award in APSIPA ASC 2024.