Adaptive Data Analysis Lab.
Associate Professor: OGAWA Tomomi
For inquiries regarding laboratory information, student placement, or courses, please contact us at the email address below.
e-mail:
to.ogawa(at)mail.dendai.ac.jp
Adaptive Signal Processing/Blind Signal Processing Demo(GitHub)
Our laboratory focuses on signal processing and its theoretical foundations, with applications spanning data science and machine learning.
Digital Signal Processing (Adaptive and Blind Signal Processing)
Data Analysis and Data Science (Medical Data Analytics, Demand Forecasting)
Media Information Processing (Video Summarization, Image and Speech Processing, Noise Reduction)
Applied Machine Learning and Real-World Deployment (ICT-based Rehabilitation Systems, Anomaly Detection for Sewer Pipe Inspection)
Our laboratory conducts research in digital signal processing and data science, with a focus on developing algorithms and analytical methods for extracting the information that truly matters from data buried in noise and redundancy. In particular, we emphasize a mathematical understanding of what information can actually be extracted and under what conditions fundamentally indistinguishable limits arise, that is, the study of non-identifiability. We pursue this research from both theoretical and implementation-oriented perspectives.
The main pillars of our work are adaptive processing for nonstationary data, signal and data analysis based on mathematical structure, and the design of algorithms that can flexibly respond to continuously changing environments. In recent years, in addition to conventional signal processing, machine learning, and statistical analysis, we have also begun to explore quantum blind signal processing, which incorporates ideas from quantum computation and quantum information. This emerging direction investigates how far information can be identified and separated under the constraints imposed by quantum measurement, and may contribute to the foundations of signal processing and information analysis in the quantum era.
Students joining the laboratory begin with a seminar on advanced linear algebra, including projection, generalized inverses, and singular value decomposition. These topics provide powerful mathematical tools for understanding signal processing, machine learning, geometric modeling, and the mathematical foundations of quantum information, and they form an important step toward building a solid command of theory and equations.
In undergraduate and graduate research, we provide an environment in which each student can work on a wide range of topics, from mathematically oriented themes to projects with potential industrial applications, depending on their interests and career goals. Current and recent research topics include music removal without access to the original source, sports video summarization, time-series analysis of sales data, anomaly detection using sewer pipe images, rehabilitation menu recommendation systems for day-service users, and improved system estimation accuracy using adaptive filters. Looking ahead, we also plan to expand into foundational research on quantum blind signal processing.
2026 Members: 1 research student, 3 graduate students, 12 undergraduate students
Hyuga HAGIWARA, Tomomi OGAWA ,and Hiroki MATSUMOTO: A consideration of data augmentation methods for improving the accuracy of music removal from environmental sounds using machine learning, Proc. of 2025 International Symposium on Intelligent Signal Processing and Communication Systems (ISPACS 2025)
Hyuga Hagiwara, Yuka Saito, Tomomi OGAWA, and Hiroki Matsumoto: Data Augmentation Methods for Music Removal from Environmental Sounds Based on Machine Learning, Proc. of The 24th International Symposium on Communications and Information Technologies (ISCIT 2025)
Hyuga Hagiwara, Tomomi OGAWA, and Hiroki Matsumoto: A Consideration on Improvement of Music Removal Accuracy from Environmental Sounds, Proc. of 2024 International Symposium on Intelligent Signal Processing and Communication Systems (ISPACS 2024)
Hyuga Hagiwara, Tomomi OGAWA, and Hiroki Matsumoto: Improvement of Music Removal Accuracy from Environmental Sounds, 2024 23rd International Symposium on Communications and Information Technologies (ISCIT)
Kiryu Mochizuki, Kohei Maeno, and Tomomi OGAWA: Granger Causality Analysis Between Feature Values Before and After the Change Point in Vehicle Sales Data, Proc. of 2024 IEEE 4th International Conference on Software Engineering and Artificial Intelligence (SEAI)
Shun Ozawa and Tomomi OGAWA: Background Music Removal Using Deep Learning, Proc. of 2023 3rd IEEE International Conference on Software Engineering a/nd Artificial Intelligence (SEAI)
Taiki Yoshida and Tomomi OGAWA: Highlight Generation Method for Game Commentary using Acoustic Features, Proc. of 2023 3rd IEEE International Conference on Software Engineering and Artificial Intelligence (SEAI)
Ryutaro Igarashi and Tomomi OGAWA: Learning Method to Detect Anomalies in Sewer Pipe Images Using Object Detection, Proc. of 2023 3rd IEEE International Conference on Software Engineering and Artificial Intelligence (SEAI)
Address: Ishizaka, Hatoyama-machi, Hiki-gun, Saitama 350-0394
Division of Information Systems and Design, School of Science and Engineering, Tokyo Denki University