Music Signal Separation Based on Supervised Nonnegative Matrix Factorization with Orthogonality and Maximum-Divergence Penalties

被引：26

作者：

Kitamura, Daichi ^{[1
]}

Saruwatari, Hiroshi ^{[1
]}

Yagi, Kosuke ^{[1
]}

Shikano, Kiyohiro ^{[1
]}

Takahashi, Yu ^{[2
]}

Kondo, Kazunobu ^{[2
]}

机构：

[1] Nara Inst Sci & Technol, Ikoma 6300192, Japan

[2] Yamaha Corp, Res & Dev, Iwata 4380192, Japan

来源：

IEICE TRANSACTIONS ON FUNDAMENTALS OF ELECTRONICS COMMUNICATIONS AND COMPUTER SCIENCES | 2014年 / E97A卷 / 05期

关键词：

music signal separation; nonnegative matrix factorization; supervised method;

D O I：

10.1587/transfun.E97.A.1113

中图分类号：

TP3 [计算技术、计算机技术];

学科分类号：

0812 ;

摘要：

In this letter, we address monaural source separation based on supervised nonnegative matrix factorization (SNMF) and propose a new penalized SNMF. Conventional SNMF often degrades the separation performance owing to the basis-sharing problem. Our penalized SNMF forces nontarget bases to become different from the target bases, which increases the separated sound quality.

引用

页码：1113 / 1118

页数：6

共 7 条

[1]

[Anonymous], 2001, TECHNICAL REPORT

[2]

Kameoka H, 2012, INT CONF ACOUST SPEE, P5365, DOI 10.1109/ICASSP.2012.6289133

[3]

Lee DD, 2001, ADV NEUR IN, V13, P556

[4]

Nakano Masahiro, 2010, Proceedings of the 2010 IEEE International Workshop on Machine Learning for Signal Processing (MLSP), P283, DOI 10.1109/MLSP.2010.5589233

[5]

Smaragdis P., 2010, LNCS, V6365, P140

[6] Performance measurement in blind audio source separation [J].

Vincent, Emmanuel ;

Gribonval, Remi ;

Févotte, Cedric .

IEEE TRANSACTIONS ON AUDIO SPEECH AND LANGUAGE PROCESSING, 2006, 14 (04) :1462-1469

[7] Monaural sound source separation by nonnegative matrix factorization with tempora continuity and sparseness criteria [J].

Virtanen, Tuomas .

IEEE TRANSACTIONS ON AUDIO SPEECH AND LANGUAGE PROCESSING, 2007, 15 (03) :1066-1074

← 1 →