In this paper, we present update rules for convolutive nonnegative matrix factorization (NMF) in which cost functions are based on the squared Euclidean distance, the Kullback-Leibler (KL) divergence and the Itakura-Saito (IS) divergence. We define an auxiliary function for each cost function and derive the update rule. We also apply this method to the single-channel signal separation in speech signals. Experimental results showed that the convergence of our KL divergence-based method was better than that in the conventional method, and our method achieved single-channel signal separation successfully.
Hiroki TANJI
Meiji University
Ryo TANAKA
Meiji University
Kyohei TABATA
Meiji University
Yoshito ISEKI
Meiji University
Takahiro MURAKAMI
Meiji University
Yoshihisa ISHIDA
Meiji University
The copyright of the original papers published on this site belongs to IEICE. Unauthorized use of the original or translated papers is prohibited. See IEICE Provisions on Copyright for details.
Copy
Hiroki TANJI, Ryo TANAKA, Kyohei TABATA, Yoshito ISEKI, Takahiro MURAKAMI, Yoshihisa ISHIDA, "Derivation of Update Rules for Convolutive NMF Based on Squared Euclidean Distance, KL Divergence, and IS Divergence" in IEICE TRANSACTIONS on Fundamentals,
vol. E97-A, no. 11, pp. 2121-2129, November 2014, doi: 10.1587/transfun.E97.A.2121.
Abstract: In this paper, we present update rules for convolutive nonnegative matrix factorization (NMF) in which cost functions are based on the squared Euclidean distance, the Kullback-Leibler (KL) divergence and the Itakura-Saito (IS) divergence. We define an auxiliary function for each cost function and derive the update rule. We also apply this method to the single-channel signal separation in speech signals. Experimental results showed that the convergence of our KL divergence-based method was better than that in the conventional method, and our method achieved single-channel signal separation successfully.
URL: https://global.ieice.org/en_transactions/fundamentals/10.1587/transfun.E97.A.2121/_p
Copy
@ARTICLE{e97-a_11_2121,
author={Hiroki TANJI, Ryo TANAKA, Kyohei TABATA, Yoshito ISEKI, Takahiro MURAKAMI, Yoshihisa ISHIDA, },
journal={IEICE TRANSACTIONS on Fundamentals},
title={Derivation of Update Rules for Convolutive NMF Based on Squared Euclidean Distance, KL Divergence, and IS Divergence},
year={2014},
volume={E97-A},
number={11},
pages={2121-2129},
abstract={In this paper, we present update rules for convolutive nonnegative matrix factorization (NMF) in which cost functions are based on the squared Euclidean distance, the Kullback-Leibler (KL) divergence and the Itakura-Saito (IS) divergence. We define an auxiliary function for each cost function and derive the update rule. We also apply this method to the single-channel signal separation in speech signals. Experimental results showed that the convergence of our KL divergence-based method was better than that in the conventional method, and our method achieved single-channel signal separation successfully.},
keywords={},
doi={10.1587/transfun.E97.A.2121},
ISSN={1745-1337},
month={November},}
Copy
TY - JOUR
TI - Derivation of Update Rules for Convolutive NMF Based on Squared Euclidean Distance, KL Divergence, and IS Divergence
T2 - IEICE TRANSACTIONS on Fundamentals
SP - 2121
EP - 2129
AU - Hiroki TANJI
AU - Ryo TANAKA
AU - Kyohei TABATA
AU - Yoshito ISEKI
AU - Takahiro MURAKAMI
AU - Yoshihisa ISHIDA
PY - 2014
DO - 10.1587/transfun.E97.A.2121
JO - IEICE TRANSACTIONS on Fundamentals
SN - 1745-1337
VL - E97-A
IS - 11
JA - IEICE TRANSACTIONS on Fundamentals
Y1 - November 2014
AB - In this paper, we present update rules for convolutive nonnegative matrix factorization (NMF) in which cost functions are based on the squared Euclidean distance, the Kullback-Leibler (KL) divergence and the Itakura-Saito (IS) divergence. We define an auxiliary function for each cost function and derive the update rule. We also apply this method to the single-channel signal separation in speech signals. Experimental results showed that the convergence of our KL divergence-based method was better than that in the conventional method, and our method achieved single-channel signal separation successfully.
ER -