Presentation 2016-09-16
An audio watermarking method using non-negative matrix factorization separated the bases of host signal and watermark signal
Harumi Murata, Akio Ogihara,
PDF Download Page PDF download Page Link
Abstract(in Japanese) (See Japanese page)
Abstract(in English) For audio signals, the sound quality of stego signal should not deteriorated as much as host signal. With current methods, high sound quality means that the difference between the host and stego signals is small. On the other hand, high sound quality means that the sound quality of stego signal is maintained as music even if another sound, except for host signal, is perceived in this paper. According to this definition, the watermarks are embedded using non-negative matrix factorization. NMF is an algorithm that factorizes a non-negative matrix that is an amplitude spectrogram of host signal into two non-negative matrices that corresponds to the spectral patterns of the host signal and the activation of each spectrum. In this paper, a basis matrix of watermark signal separation is different from that of host signal separation. The watermarks are embedded by the magnitude comparison of activation matrix coefficients of watermark signal corresponding to root and dominant notes. It is considered that the components of watermark signal can be only extracted from the stego signal, because the basis matrices of host and watermark signal are different. We conduct an evaluation of the tolerance against attacks and an objective evaluation of sound quality, and confirm the validity of proposed method.
Keyword(in Japanese) (See Japanese page)
Keyword(in English) Audio Watermarking / Non-negative Matrix Factorization
Paper # LOIS2016-26,IE2016-63,EMM2016-52
Date of Issue 2016-09-08 (LOIS, IE, EMM)

Conference Information
Committee ITE-ME / IE / EMM / LOIS / IEE-CMN
Conference Date 2016/9/15(2days)
Place (in Japanese) (See Japanese page)
Place (in English) Aichi Prefectural University
Topics (in Japanese) (See Japanese page)
Topics (in English)
Chair Miki Haseyama(Hokkaido Univ.) / Seishi Takamura(NTT) / Akinori Ito(Tohoku Univ.) / Hiroyuki Nishi(Sojo Univ.) / Yoshimi Fukumura(Nagaoka Univ. of Tech.)
Vice Chair Norio Tagawa(Tokyo Metropolitan Univ.) / Takayuki Hamamoto(Tokyo Univ. of Science) / Atsuro Ichigaya(NHK) / Masaki Kawamura(Yamaguchi Univ.) / Hirohisa Hioki(Kyoto Univ.) / Tomohiro Yamada(NTT)
Secretary Norio Tagawa(NEC) / Takayuki Hamamoto(NTT) / Atsuro Ichigaya(Chiba Inst. of Tech.) / Masaki Kawamura(Nagasaki Univ.) / Hirohisa Hioki(Osaka Pref. Univ.) / Tomohiro Yamada(NTT) / (Nagasaki Univ.)
Assistant / Kei Kawamura(KDDI R&D Labs.) / Keita Takahashi(Nagoya Univ.) / Rui Shogenji(Shizuoka Univ.) / Masaaki Fujiyoshi(Tokyo Metropolitan Univ.) / Yukihiro Nakamura(NTT) / Keisuke Utsu(Tokai Univ.) / Takeshi YAGI(NTT)

Paper Information
Registration To Technical Group on Media Engineering / Technical Committee on Image Engineering / Technical Committee on Enriched MultiMedia / Technical Committee on Life Intelligence and Office Information Systems / Technical Group on Communications
Language JPN
Title (in Japanese) (See Japanese page)
Sub Title (in Japanese) (See Japanese page)
Title (in English) An audio watermarking method using non-negative matrix factorization separated the bases of host signal and watermark signal
Sub Title (in English)
Keyword(1) Audio Watermarking
Keyword(2) Non-negative Matrix Factorization
1st Author's Name Harumi Murata
1st Author's Affiliation Chukyo University(Chukyo Univ.)
2nd Author's Name Akio Ogihara
2nd Author's Affiliation Kindai University(Kindai Univ.)
Date 2016-09-16
Paper # LOIS2016-26,IE2016-63,EMM2016-52
Volume (vol) vol.116
Number (no) LOIS-220,IE-221,EMM-222
Page pp.pp.61-66(LOIS), pp.61-66(IE), pp.61-66(EMM),
#Pages 6
Date of Issue 2016-09-08 (LOIS, IE, EMM)