Independent Low-Rank Matrix Analysis Based on Generalized Kullback-Leibler Divergence

• MOGAMI Shinichi

  University of Tokyo

• MITSUI Yoshiki

  University of Tokyo

• TAKAMUNE Norihiro

  University of Tokyo

• KITAMURA Daichi

  National Institute of Technology, Kagawa College

• SARUWATARI Hiroshi

  University of Tokyo

• TAKAHASHI Yu

  Yamaha Corporation

• KONDO Kazunobu

  Yamaha Corporation

• NAKAJIMA Hiroaki

  Yamaha Corporation

• KAMEOKA Hirokazu

  Nippon Telegraph and Telephone Corporation

Abstract

In this letter, we propose a new blind source separation method, independent low-rank matrix analysis based on generalized Kullback-Leibler divergence. This method assumes a time-frequency-varying complex Poisson distribution as the source generative model, which yields convex optimization in the spectrogram estimation. The experimental evaluation confirms the proposed method's efficacy.

Journal

• IEICE Transactions on Fundamentals of Electronics, Communications and Computer Sciences

  IEICE Transactions on Fundamentals of Electronics, Communications and Computer Sciences E102.A (2), 458-463, 2019-02-01

  The Institute of Electronics, Information and Communication Engineers

Keywords

• blind source separation
• nonnegative matrix factorization
• Poisson distribution
• Kullback-Leibler divergence

Details

• CRID: 1390845713049723008
• NII Article ID: 130007585927
• DOI: 10.1587/transfun.e102.a.458
• ISSN: 17451337
• ISSN: 09168508
• Web Site: https://www.jstage.jst.go.jp/article/transfun/E102.A/2/E102.A_458/_pdf
• Text Lang: en
• Data Source:
  • JaLC
  • Crossref
  • CiNii Articles
  • KAKEN
• Abstract License Flag: Disallowed