Committee |
Date Time |
Place |
Paper Title / Authors |
Abstract |
Paper # |
SIP, SP, EA, IPSJ-SLP [detail] |
2024-02-29 15:10 |
Okinawa |
(Primary: On-site, Secondary: Online) |
Cramér-Rao Lower Bound for Parameter Estimation from Observation with Irreversible Saturation Effects Natsuki Ueno, Hirokazu Kameoka (NTT) EA2023-84 SIP2023-131 SP2023-66 |
For probabilistic models involving irreversible saturation effects, we provide extended definitions of the score functio... [more] |
EA2023-84 SIP2023-131 SP2023-66 pp.139-144 |
SIP, SP, EA, IPSJ-SLP [detail] |
2024-03-01 09:30 |
Okinawa |
(Primary: On-site, Secondary: Online) |
SELECTING N-LOWEST SCORES FOR TRAINING MOS PREDICTION MODELS Yuto Kondo, Hirokazu Kameoka, Kou Tanaka, Takuhiro Kaneko (NTT) EA2023-94 SIP2023-141 SP2023-76 |
Automatic speech quality assessment (SQA) is a task to evaluate the quality of speech samples without resorting to time-... [more] |
EA2023-94 SIP2023-141 SP2023-76 pp.196-201 |
EA |
2019-12-13 13:00 |
Fukuoka |
Kyushu Inst. Tech. |
Speaker-independent source separation with multichannel variational autoencoder Li Li (Univ. Tsukuba), Hirokazu Kameoka (NTT), Shota Inoue, Shoji Makino (Univ. Tsukuba) EA2019-77 |
The multichannel variational autoencoder method (MVAE) is a recently proposed determined source separation method, which... [more] |
EA2019-77 pp.79-84 |
SP |
2019-08-28 13:30 |
Kyoto |
Kyoto Univ. |
WaveCycleGAN2: Neural Waveform Post-Filter For High-Quality Speech Generation Kou Tanaka, Hirokazu Kameoka, Takuhiro Kaneko, Nobukatsu Hojo (NTT) SP2019-9 |
[more] |
SP2019-9 pp.1-6 |
SP |
2019-08-28 13:55 |
Kyoto |
Kyoto Univ. |
Sequence-to-Sequence Voice Conversion Using Context Preservation Mechanism Kou Tanaka, Hirokazu Kameoka, Takuhiro Kaneko, Nobukatsu Hojo (NTT) SP2019-10 |
[more] |
SP2019-10 pp.7-12 |
EA, SIP, SP |
2019-03-14 13:30 |
Nagasaki |
i+Land nagasaki (Nagasaki-shi) |
[Poster Presentation]
CWT spectral loss for training a DNN-based speech waveform model Shinji Takaki (NII), Hirokazu Kameoka (NTT), Junichi Yamagishi (NII) EA2018-121 SIP2018-127 SP2018-83 |
[more] |
EA2018-121 SIP2018-127 SP2018-83 pp.131-135 |
EA, SIP, SP |
2019-03-14 13:30 |
Nagasaki |
i+Land nagasaki (Nagasaki-shi) |
[Poster Presentation]
A robust algorithm of phase recovery for speech enhancement Dongxiao Wang, Koichi Shinoda (TokyoTech), Hirokazu Kameoka (NTT) EA2018-122 SIP2018-128 SP2018-84 |
[more] |
EA2018-122 SIP2018-128 SP2018-84 pp.137-142 |
EA, SIP, SP |
2019-03-15 13:30 |
Nagasaki |
i+Land nagasaki (Nagasaki-shi) |
[Poster Presentation]
An Evaluation of Underdetermined Source Separation Based on Multichannel Variational Autoencoder Shogo Seki (Nagoya Univ.), Hirokazu Kameoka (NTT), Li Li (Univ. Tsukuba), Tomoki Toda, Kazuya Takeda (Nagoya Univ.) EA2018-154 SIP2018-160 SP2018-116 |
This paper deals with a multichannel audio source separation problem under underdetermined conditions. Multichannel Non-... [more] |
EA2018-154 SIP2018-160 SP2018-116 pp.323-328 |
SIP, EA, SP, MI (Joint) [detail] |
2018-03-20 09:00 |
Okinawa |
|
[Poster Presentation]
Nonnegative Matrix Factorization for Determined Multichannel Systems under Reverberant Environments Hideaki Kagami (Keio Univ.), Hirokazu Kameoka (NTT), Masahiro Yukawa (Keio Univ.) EA2017-153 SIP2017-162 SP2017-136 |
[more] |
EA2017-153 SIP2017-162 SP2017-136 pp.281-286 |
EA, ASJ-H |
2017-10-22 09:00 |
Toyama |
Ushidake-Onsen |
[Invited Talk]
Blind source separation based on independent low-rank matrix analysis Daichi Kitamura (UT), Nobutaka Ono (NII), Hiroshi Sawada, Hirokazu Kameoka (NTT), Hiroshi Saruwatari (UT) EA2017-56 |
In this paper, we propose a new effective algorithm for blind source separation problem (BSS) called independent low-ran... [more] |
EA2017-56 pp.73-80 |
PRMU, SP |
2017-06-22 14:45 |
Miyagi |
|
Postfiltering of STFT Spectrograms Based on Generative Adversarial Networks Takuhiro Kaneko (NTT), Shinji Takaki (NII), Hirokazu Kameoka (NTT), Junichi Yamagishi (NII) PRMU2017-28 SP2017-4 |
This paper presents postfiltering of short-term Fourier transform (STFT) spectrograms based on Generative Adversarial Ne... [more] |
PRMU2017-28 SP2017-4 pp.17-22 |
SP, SIP, EA |
2017-03-01 09:45 |
Okinawa |
Okinawa Industry Support Center |
Nonaudible murmur enhancement based on non-negative tensor factorization with segment feature regularization in noisy environments Yusuke Tajiri (Nagoya Univ.), Hirokazu Kameoka (NTT), Tomoki Toda (Nagoya Univ.) EA2016-83 SIP2016-138 SP2016-78 |
Towards the development of silent speech communication, there has been studied a statistical approach to enhancing nonau... [more] |
EA2016-83 SIP2016-138 SP2016-78 pp.7-12 |
SP, SIP, EA |
2017-03-01 10:50 |
Okinawa |
Okinawa Industry Support Center |
Missing Component Restoration for Speech Spectrogram Based on Time-domain Signal Estimation Shogo Seki (Nagoya Univ.), Hirokazu Kameoka (NTT), Tomoki Toda, Kazuya Takeda (Nagoya Univ.) EA2016-85 SIP2016-140 SP2016-80 |
This study proposes a missing component restoration method for time-frequency masked speech spectrogram based on time-do... [more] |
EA2016-85 SIP2016-140 SP2016-80 pp.19-24 |
SP, SIP, EA |
2017-03-02 09:00 |
Okinawa |
Okinawa Industry Support Center |
[Poster Presentation]
Acoustic-to-articulatory inversion mapping with variational latent trajectory Gaussian mixture model Patrick Lumban Tobing (Nagoya Univ.), Hirokazu Kameoka (NTT), Tomoki Toda (Nagoya Univ.) EA2016-134 SIP2016-189 SP2016-129 |
[more] |
EA2016-134 SIP2016-189 SP2016-129 pp.291-296 |
SP, SIP, EA |
2017-03-02 12:45 |
Okinawa |
Okinawa Industry Support Center |
Non-native speech conversion with consistency-aware recursive network and generative adversarial network Keisuke Oyamada (Univ. of Tsukuba), Hirokazu Kameoka, Takuhiro Kaneko (NTT), Hiroyasu Ando (Univ. of Tsukuba), Kaoru Hiramatsu, Kunio Kashino (NTT) EA2016-139 SIP2016-194 SP2016-134 |
This paper deals with the problem of automatically modifying the pronunciation of non-native speech.
Since the pronunci... [more] |
EA2016-139 SIP2016-194 SP2016-134 pp.315-320 |
SP, IPSJ-SLP, NLC, IPSJ-NL (Joint) [detail] |
2016-12-20 15:10 |
Tokyo |
NTT Musashino R&D |
[Poster Presentation]
Fast algorithm for statistical phrase/accent command estimation based on generative model incorporating spectral features Ryotaro Sato (The Univ. of Tokyo), Hirokazu Kameoka, Kunio Kashino (NTT) SP2016-56 |
On the basis of the Fujisaki model, we propose a fast algorithm for estimating the model parameters, namely, the timings... [more] |
SP2016-56 pp.43-48 |
SP, IPSJ-SLP, NLC, IPSJ-NL (Joint) [detail] |
2016-12-20 16:40 |
Tokyo |
NTT Musashino R&D |
Generative Adversarial Network-based Postfiltering for Statistical Parametric Speech Synthesis Takuhiro Kaneko, Hirokazu Kameoka, Nobukatsu Hojo, Yusuke Ijima, Kaoru Hiramatsu, Kunio Kashino (NTT) SP2016-61 |
In the field of speech synthesis, statistical parametric speech synthesis has been widely used due to the flexibility an... [more] |
SP2016-61 pp.89-94 |
SP |
2016-08-24 16:15 |
Kyoto |
ACCMS, Kyoto Univ. |
[Poster Presentation]
Joint Enhancement of Spectral and Cepstral Sequences of Noisy Speech Li Li (Univ.Tsukuba), Hirokazu Kameoka, Takuya Higuchi (NTT), Hiroshi Saruwatari (Univ.Tokyo), Shoji Makino (Univ.Tsukuba) SP2016-32 |
While spectral domain speech enhancement algorithms using non-negative matrix factorization (NMF) are powerful in terms ... [more] |
SP2016-32 pp.29-32 |
EA, SP, SIP |
2016-03-28 13:15 |
Oita |
Beppu International Convention Center B-ConPlaza |
[Poster Presentation]
Super-Resolution Vocal Tract Spectrum Estimation with Missing Data Imputation Using Non-Negative Matrix Factorization Tomohiko Nakamura (Todai), Hirokazu Kameoka (Todai/NTT) EA2015-83 SIP2015-132 SP2015-111 |
This report addresses the problem of estimating vocal tract spectra from speech signals. Spectra of speech signals can b... [more] |
EA2015-83 SIP2015-132 SP2015-111 pp.99-104 |
EA, SP, SIP |
2016-03-28 13:15 |
Oita |
Beppu International Convention Center B-ConPlaza |
[Poster Presentation]
Nonaudible murmur enhancement based on non-negative tensor factorization of air- and body-conducted signals in real environments Yusuke Tajiri (NAIST), Hirokazu Kameoka (NTT), Tomoki Toda (Nagoya Univ./NAIST), Satoshi Nakamura (NAIST) EA2015-86 SIP2015-135 SP2015-114 |
Nonaudible murmur (NAM) recorded with a special body-conductive microphone called NAM microphone is one of the promising... [more] |
EA2015-86 SIP2015-135 SP2015-114 pp.117-122 |