日本語
College of Information Science and Engineering  /
Department of Information Science and Engineering

 (Male)
 AKIRA   HIRABAYASHI  Professor

■Concurrent affiliation
Graduate School of Information Science and Engineering
■Graduate school/University/other
03/1993  Tokyo Institute of Technology  Faculty of Engineering  Department of Computer Science  Graduated
03/1995  Tokyo Institute of Technology  Graduate School of Information Science and Engineering  Department of Computer Science  Master's course  Completed
■Academic degrees
Doctor of Engineering (12/1999 Tokyo Institute of Technology)  
■Career history
04/01/1995-08/31/2000  Tokyo Institute of Technology, Research Associate
09/01/2000-09/30/2002  Yamaguchi University, Lecturer
10/31/2002-03/31/2013  Yamaguchi University, Associate Professor
09/21/2004-03/20/2005  Swiss Federal Institute of Technology, Invited Professor
08/21/2009-02/16/2010  Imperial College London, Visiting Associate Professor
■Committee history
05/2005-04/2007  Institute of Electrical and Electronics Engineers (IEEE), Hiroshima Section  Committee Member
05/2009-  Institute of Electronics, Information and Communication Engineers  Committee Member of Signal Processing Society
■Academic society memberships
Institute of Electrical and Electronics Engineers (IEEE)  
Institute of Electronics, Information and Communication Engineers  
Society of Instrument and Control Engineers (SICE)  
■Subject of research
Sparse sampling theory and applications to vehicular and medical signals
Line edge extraction from digital images
Compressed sensing algorithm for medical imaging
■Research summary
Signal and image sensing theory and its applications

 Sensing is the starting point of information processing. In particular, sampling enables us to handle analog signals in digital computers. The standard approach for sampling is band-limitedness presented by Shannon or Someya. This principle is not always appropriate, especially for signals that have sparse representation in a certain domain. For such signals, data acquisition at low sampling frequency is possible (sparse sampling). Exploiting parametric representations for sparse signals, we are developing precise and fast recovery algorithms based on maximum-likelihood and maximum a posteriori estimations. This problem can be regarded as an inverse problem. This includes image restoration and super-resolution, which we are also tackling. We are further doing research about brain signal analysis.
■Research keywords
Signal Processing, Image Processing, Sensing, and Sampling 
■Research activities   (Even top three results are displayed. In View details, all results for public presentation are displayed.)

Books
"Sampling and recovery of continuously-defined sparse signals and its applications,”Chapter 8 in Subspace Methods for Pattern Recognition in Intelligent Environment, Y.W. Chen, ed.,  A. Hirabayashi  Chapter 8  2013
“信号の標本化,”電子情報通信学会知識ベース「知識の森」,
原島博(編)  平林晃  電気情報通信学会  第1 群, 第5 編, 第3 章, 全14 ページ  2010
View details...

Papers
Complex approximate message passing algorithm for two-dimensional compressed sensing  A. Hirabayashi, J. Sugimoto and K. Mimura  IEICE Transactions on Fundamentals of Electronics, Communications and Computer Sciences  vol. E96-A, pp. 2391–2397  2013
Sampling signals with finite rate of innovation and recovery by maximum likelihood estimation  A.Hirabayashi, Y. Hironaga and L. Condat  IEICE Transactions on Fundamentals of Electronics, Communications and Computer Sciences  vol. E96-A, pp.1972-1979  2013
連続信号のスパースサンプリング―ナイキストの壁を越えて  平林晃  電子情報通信学会誌  vol. 96, pp. 269-273  2013
View details...

Research presentations
Recovery of nonuniform Dirac pulses from noisy linear measurements  Proceedings of the International Conference on Acoustics, Speech, and Signal Processing (ICASSP 2013)  2013
Approximate message passing algorithm for complex separable compressed imaging  Proceedings of the 2013 Asia-Pacific Signal and Information
Processing Association Annual Summit and Conference (APSIPA ASC 2013)  2013
Robust spike train recovery from noisy data by structured low rank approximation  Proceedings of the 2013 Sampling Theory and Applications (SampTA 2013)  2013
View details...

Grants-in-Aid for Scientific Research (KAKENHI)
Link to Grants-in-Aid for Scientific Research -KAKENHI-

Competitive grants, etc. (exc. KAKENHI)
圧縮センシングに基づく磁気共鳴画像(MRI)の高速撮像法の開発  立石科学技術財団研究助成(B)  04/2012  03/2013  Main representative
ハフ変換を越える高精度直線抽出手法の開発  JST 研究成果最適展開支援事業(A-STEP)「探索タイプ」  09/2010  03/2011  Main representative
超解像処理のための画像変換パラメータの推定法に関する研究  科学技術振興機構(JST)平成18年度「シーズ発掘試験」  09/2006  02/2007  Main representative
View details...
■Teaching experience   (Even top three results are displayed. In View details, all results for public presentation are displayed.)

Courses taught
2016  Digital Signal Processing  Lecture
2016  Programming Seminar 2  Seminar
2016  Media Project Seminar 2  Seminar
View details...

Teaching achievements
東山高校 模擬講義 映像メディア技術入門  05/2014-05/2014
View details...
■Message from researcher
Sensing and sampling in signal and image processing
 We are doing research about sensing and sampling related to signal and image processing. The main topic is sparse sensing, which can be applied to, for example, communication traffic reduction in sensor networks or shape recognition in digital images. The significant feature is that we directly acquire the data, which is the same as, or as much as, the compressed data after mega-sensing, such as JPEG image compression. Thus, we can implement small and fast sensors.
■Research keywords(on a multiple-choice system)
Mathematical informatics
Perceptual information processing