Complex NMF with the generalized Kullback-Leibler divergence

Hirokazu Kameoka, Hideaki Kagami, Masahiro Yukawa

Research output: Chapter in Book/Report/Conference proceedingConference contribution

6 Citations (Scopus)

Abstract

We previously introduced a phase-aware variant of the non-negative matrix factorization (NMF) approach for audio source separation, which we call the 'Complex NMF (CNMF).' This approach makes it possible to realize NMF-like signal decompositions in the complex time-frequency domain. One limitation of the CNMF framework is that the divergence measure is limited to only the Euclidean distance. Some previous studies have revealed that for source separation tasks with NMF, the generalized Kullback-Leibler (KL) divergence tends to yield higher accuracy than when using other divergence measures. This motivated us to believe that CNMF could achieve even greater source separation accuracy if we could derive an algorithm for a KL divergence counterpart of CNMF. In this paper, we start by defining the notion of the 'dual' form of the CNMF formulation, derived from the original Euclidean CNMF, and show that a KL divergence counterpart of CNMF can be developed based on this dual formulation. We call this 'KL-CNMF'. We further derive a convergence-guaranteed iterative algorithm for KL-CNMF based on a majorization-minimization scheme. The source separation experiments revealed that the proposed KL-CNMF yielded higher accuracy than the Euclidean CNMF and NMF with varying divergences.

Original languageEnglish
Title of host publication2017 IEEE International Conference on Acoustics, Speech, and Signal Processing, ICASSP 2017 - Proceedings
PublisherInstitute of Electrical and Electronics Engineers Inc.
Pages56-60
Number of pages5
ISBN (Electronic)9781509041176
DOIs
Publication statusPublished - 2017 Jun 16
Event2017 IEEE International Conference on Acoustics, Speech, and Signal Processing, ICASSP 2017 - New Orleans, United States
Duration: 2017 Mar 52017 Mar 9

Other

Other2017 IEEE International Conference on Acoustics, Speech, and Signal Processing, ICASSP 2017
CountryUnited States
CityNew Orleans
Period17/3/517/3/9

Fingerprint

Source separation
Factorization
Decomposition
Experiments

Keywords

  • Audio source separation
  • Complex NMF
  • generalized Kullback-Leibler (KL) divergence
  • non-negative matrix factorization (NMF)

ASJC Scopus subject areas

  • Software
  • Signal Processing
  • Electrical and Electronic Engineering

Cite this

Kameoka, H., Kagami, H., & Yukawa, M. (2017). Complex NMF with the generalized Kullback-Leibler divergence. In 2017 IEEE International Conference on Acoustics, Speech, and Signal Processing, ICASSP 2017 - Proceedings (pp. 56-60). [7952117] Institute of Electrical and Electronics Engineers Inc.. https://doi.org/10.1109/ICASSP.2017.7952117

Complex NMF with the generalized Kullback-Leibler divergence. / Kameoka, Hirokazu; Kagami, Hideaki; Yukawa, Masahiro.

2017 IEEE International Conference on Acoustics, Speech, and Signal Processing, ICASSP 2017 - Proceedings. Institute of Electrical and Electronics Engineers Inc., 2017. p. 56-60 7952117.

Research output: Chapter in Book/Report/Conference proceedingConference contribution

Kameoka, H, Kagami, H & Yukawa, M 2017, Complex NMF with the generalized Kullback-Leibler divergence. in 2017 IEEE International Conference on Acoustics, Speech, and Signal Processing, ICASSP 2017 - Proceedings., 7952117, Institute of Electrical and Electronics Engineers Inc., pp. 56-60, 2017 IEEE International Conference on Acoustics, Speech, and Signal Processing, ICASSP 2017, New Orleans, United States, 17/3/5. https://doi.org/10.1109/ICASSP.2017.7952117
Kameoka H, Kagami H, Yukawa M. Complex NMF with the generalized Kullback-Leibler divergence. In 2017 IEEE International Conference on Acoustics, Speech, and Signal Processing, ICASSP 2017 - Proceedings. Institute of Electrical and Electronics Engineers Inc. 2017. p. 56-60. 7952117 https://doi.org/10.1109/ICASSP.2017.7952117
Kameoka, Hirokazu ; Kagami, Hideaki ; Yukawa, Masahiro. / Complex NMF with the generalized Kullback-Leibler divergence. 2017 IEEE International Conference on Acoustics, Speech, and Signal Processing, ICASSP 2017 - Proceedings. Institute of Electrical and Electronics Engineers Inc., 2017. pp. 56-60
@inproceedings{f876dfef657e4e2a8d112b6fe5a1d1a9,
title = "Complex NMF with the generalized Kullback-Leibler divergence",
abstract = "We previously introduced a phase-aware variant of the non-negative matrix factorization (NMF) approach for audio source separation, which we call the 'Complex NMF (CNMF).' This approach makes it possible to realize NMF-like signal decompositions in the complex time-frequency domain. One limitation of the CNMF framework is that the divergence measure is limited to only the Euclidean distance. Some previous studies have revealed that for source separation tasks with NMF, the generalized Kullback-Leibler (KL) divergence tends to yield higher accuracy than when using other divergence measures. This motivated us to believe that CNMF could achieve even greater source separation accuracy if we could derive an algorithm for a KL divergence counterpart of CNMF. In this paper, we start by defining the notion of the 'dual' form of the CNMF formulation, derived from the original Euclidean CNMF, and show that a KL divergence counterpart of CNMF can be developed based on this dual formulation. We call this 'KL-CNMF'. We further derive a convergence-guaranteed iterative algorithm for KL-CNMF based on a majorization-minimization scheme. The source separation experiments revealed that the proposed KL-CNMF yielded higher accuracy than the Euclidean CNMF and NMF with varying divergences.",
keywords = "Audio source separation, Complex NMF, generalized Kullback-Leibler (KL) divergence, non-negative matrix factorization (NMF)",
author = "Hirokazu Kameoka and Hideaki Kagami and Masahiro Yukawa",
year = "2017",
month = "6",
day = "16",
doi = "10.1109/ICASSP.2017.7952117",
language = "English",
pages = "56--60",
booktitle = "2017 IEEE International Conference on Acoustics, Speech, and Signal Processing, ICASSP 2017 - Proceedings",
publisher = "Institute of Electrical and Electronics Engineers Inc.",

}

TY - GEN

T1 - Complex NMF with the generalized Kullback-Leibler divergence

AU - Kameoka, Hirokazu

AU - Kagami, Hideaki

AU - Yukawa, Masahiro

PY - 2017/6/16

Y1 - 2017/6/16

N2 - We previously introduced a phase-aware variant of the non-negative matrix factorization (NMF) approach for audio source separation, which we call the 'Complex NMF (CNMF).' This approach makes it possible to realize NMF-like signal decompositions in the complex time-frequency domain. One limitation of the CNMF framework is that the divergence measure is limited to only the Euclidean distance. Some previous studies have revealed that for source separation tasks with NMF, the generalized Kullback-Leibler (KL) divergence tends to yield higher accuracy than when using other divergence measures. This motivated us to believe that CNMF could achieve even greater source separation accuracy if we could derive an algorithm for a KL divergence counterpart of CNMF. In this paper, we start by defining the notion of the 'dual' form of the CNMF formulation, derived from the original Euclidean CNMF, and show that a KL divergence counterpart of CNMF can be developed based on this dual formulation. We call this 'KL-CNMF'. We further derive a convergence-guaranteed iterative algorithm for KL-CNMF based on a majorization-minimization scheme. The source separation experiments revealed that the proposed KL-CNMF yielded higher accuracy than the Euclidean CNMF and NMF with varying divergences.

AB - We previously introduced a phase-aware variant of the non-negative matrix factorization (NMF) approach for audio source separation, which we call the 'Complex NMF (CNMF).' This approach makes it possible to realize NMF-like signal decompositions in the complex time-frequency domain. One limitation of the CNMF framework is that the divergence measure is limited to only the Euclidean distance. Some previous studies have revealed that for source separation tasks with NMF, the generalized Kullback-Leibler (KL) divergence tends to yield higher accuracy than when using other divergence measures. This motivated us to believe that CNMF could achieve even greater source separation accuracy if we could derive an algorithm for a KL divergence counterpart of CNMF. In this paper, we start by defining the notion of the 'dual' form of the CNMF formulation, derived from the original Euclidean CNMF, and show that a KL divergence counterpart of CNMF can be developed based on this dual formulation. We call this 'KL-CNMF'. We further derive a convergence-guaranteed iterative algorithm for KL-CNMF based on a majorization-minimization scheme. The source separation experiments revealed that the proposed KL-CNMF yielded higher accuracy than the Euclidean CNMF and NMF with varying divergences.

KW - Audio source separation

KW - Complex NMF

KW - generalized Kullback-Leibler (KL) divergence

KW - non-negative matrix factorization (NMF)

UR - http://www.scopus.com/inward/record.url?scp=85023775492&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=85023775492&partnerID=8YFLogxK

U2 - 10.1109/ICASSP.2017.7952117

DO - 10.1109/ICASSP.2017.7952117

M3 - Conference contribution

AN - SCOPUS:85023775492

SP - 56

EP - 60

BT - 2017 IEEE International Conference on Acoustics, Speech, and Signal Processing, ICASSP 2017 - Proceedings

PB - Institute of Electrical and Electronics Engineers Inc.

ER -