Committee |
Date Time |
Place |
Paper Title / Authors |
Abstract |
Paper # |
SIP |
2022-08-26 11:42 |
Okinawa |
Nobumoto Ohama Memorial Hall (Ishigaki Island) (Primary: On-site, Secondary: Online) |
Study on Relationship Between Speakers' Physiological Structure and Acoustic Speech Signals: Data-Driven Study Based on Frequency-Wise Attentional Neural Network Li Kai (JAIST), Xugang Lu (NICT), Masato Akagi, Jianwu Dang (JAIST), Sheng Li (NICT), Unoki Masashi (JAIST) SIP2022-68 |
Quantitatively revealing the relationship between speakers’ physiological structure and acoustic speech signals by consi... [more] |
SIP2022-68 pp.97-102 |
SP, IPSJ-MUS, IPSJ-SLP [detail] |
2022-06-17 13:00 |
Online |
Online |
A Study of Speech Recognition Result Correction Using BERT for Speech Translation Tadashi Ogura, Masakiyo Fujimoto, Peng Shen, Xugang Lu, Hisashi Kawai (NICT) SP2022-4 |
Speech translation (ST) technology consists of automatic speech recognition (ASR) and machine translation technologies. ... [more] |
SP2022-4 pp.10-13 |
SP |
2016-08-25 13:35 |
Kyoto |
ACCMS, Kyoto Univ. |
Diversity-driven Semi-supervised Ensemble DNN Acoustic Model Training Sheng Li (Kyoto Univ.), Xugang Lu (NICT), Shinsuke Sakai, Tatsuya Kawahara (Kyoto Univ.) SP2016-40 |
We focus on effective training DNN (Deep Neural Network) acoustic models for Chinese spoken lectures with only limited l... [more] |
SP2016-40 pp.71-76 |
PRMU |
2015-12-21 09:30 |
Nagano |
|
Evaluation of Automatic Prototype-Model Size Optimization in Large Geometric Margin Minimum Classification Error Training Masahiro Ogino (Doshisha Univ.), Hideyuki Watanabe (NICT), Shigeru Katagiri, Miho Osaki (Doshisha Univ.), Xugang Lu, Hisashi Kawai (NICT) PRMU2015-100 |
To develop a method for nding an appropriate class model size, which leads to accurate classication over unseen patter... [more] |
PRMU2015-100 pp.1-6 |
SP, IPSJ-SLP (Joint) |
2015-07-16 17:20 |
Nagano |
Katakura Suwako Hotel |
Experimental evaluation of network size effect in speaker adaptive trained DNNs embedding linear transformation networks Tsubasa Ochiai (Doshisha Univ./NICT), Shigeki Matsuda (Doshisha Univ.), Hideyuki Watanabe, Xugang Lu, Hisashi Kawai (NICT), Shigeru Katagiri (Doshisha Univ.) SP2015-41 |
Recently we proposed a novel speaker adaptation method that applied the Speaker Adaptive Training
(SAT) concept to DNN-... [more] |
SP2015-41 pp.31-36 |
PRMU, IPSJ-CVIM, MVE [detail] |
2015-01-23 09:50 |
Nara |
|
Analysis of Minimum Classification Error Training using Bit-String-Based Genetic Algorithms Hiroto Togoe (Doshisha Univ.), Hideyuki Watanabe (NICT), Shigeru Katagiri (Doshisha Univ.), Xugang Lu, Chiori Hori (NICT), Miho Ohsaki (Doshisha Univ.) PRMU2014-100 MVE2014-62 |
Minimum Classification Error (MCE) training using gradient-descent-based loss minimization does not guarantee a global m... [more] |
PRMU2014-100 MVE2014-62 pp.171-176 |
PRMU, IPSJ-CVIM, MVE [detail] |
2015-01-23 10:15 |
Nara |
|
Relation between Data Grouping and Robustness to Unseen Data in Large Geometric Margin Minimum Classification Error Training Hiroyuki Shiraishi (Doshisha Univ), Hideyuki Watanabe (NICT), Shigeru Katagiri (Doshisha Univ), Xugang Lu, Chiori Hori (NICT), Miho Ohsaki (Doshisha Univ) PRMU2014-101 MVE2014-63 |
To develop a pattern classifier that is robust to unseen pattern samples, classifier parameters have been conventionally... [more] |
PRMU2014-101 MVE2014-63 pp.177-182 |
EA |
2014-12-12 15:40 |
Ishikawa |
Satellite Plaza of Kanazawa University |
[Poster Presentation]
Study on signal to noise ratio estimation based on optimal design of subband voice activity detection Shota Morita (JAIST), Xugang Lu (NICT), Masashi Unoki (JAIST) EA2014-46 |
Estimation of signal to noise ratio (SNR) of speech plays an important role of noise reduction and speech intelligibilit... [more] |
EA2014-46 pp.37-42 |
SP, IPSJ-MUS |
2014-05-25 11:30 |
Tokyo |
|
Modulation transfer function based robust method of voice activity detection for noisy reverberant environments
-- Utilization of subband SNR estimation -- Shota Morita, Masashi Unoki (JAIST), Xugang Lu (NICT), Masato Akagi (JAIST) SP2014-41 |
Most of the current voice activity detection (VAD) algorithms deal with clean speech or additive noisy speech. However, ... [more] |
SP2014-41 pp.383-388 |
EA, EMM |
2012-11-16 11:45 |
Oita |
OITA Univ. |
Unified denoising and dereverberation method used in restoration of MTF-based power envelope Masashi Unoki, Shota Morita (JAIST), Xugang Lu (NICT) EA2012-86 EMM2012-68 |
Recent methods of speech enhancement have been proposed to suppress the effects of background noise and reverberation. T... [more] |
EA2012-86 EMM2012-68 pp.29-34 |
EA, SP, SIP |
2012-05-24 11:10 |
Osaka |
Osaka Univ. Nakanoshima Center |
Voice activity detection in MTF-based power envelope restoration Masashi Unoki (JAIST), Xugang Lu (NICT), Rico Petrick (TUD), Shota Morita, Masato Akagi (JAIST), Ruediger Hoffmann (TUD) EA2012-2 SIP2012-2 SP2012-2 |
This paper reports comparative evaluations of conventional voice activity detection (VAD) methods in reverberant environ... [more] |
EA2012-2 SIP2012-2 SP2012-2 pp.7-12 |
EA, SIP, SP |
2011-05-12 13:00 |
Osaka |
Ritsumeikan Univ. |
Study on the power envelope restoration based on the MTF concept and its application to ASR systems in noisy reverberant environments Shota Morita (JAIST), Xugang Lu (NICT), Masashi Unoki, Masato Akagi (JAIST), Ruediger Hoffmann (TUD) EA2011-7 SIP2011-7 SP2011-7 |
We previously proposed a method for restoring the speech power envelope from noisy reverberant speech based on a simple ... [more] |
EA2011-7 SIP2011-7 SP2011-7 pp.37-42 |
SP |
2008-07-17 - 2008-07-19 |
Iwate |
Iwate Prefectural Univ. |
Robust Front End Processing for Speech Recognition in Reverberant Environments: Utilization of Speech Properties Rico Petric (Dresden Univ. of Tech), Xugang Lu, Masashi Unoki, Masato Akagi (JAIST), Ruediger Hoffmann (Dresden Univ. of Tech) SP2008-44 |
This paper proposes two methods of robust automatic speech recognition (ASR) in reverberant environments. Unlike other m... [more] |
SP2008-44 pp.7-12 |
SP |
2008-06-27 - 2008-06-28 |
Hokkaido |
|
Comprehensive evaluations of BC speech restoration method based on linear prediction Masashi Unoki, Xugang Lu, Thang tat Vu, Kota Kinugasa, Masato Akagi (JAIST) SP2008-24 |
Bone-conducted (BC) speech can be used instead of air-conducted (AC) speech for speech communication and speech recognit... [more] |
SP2008-24 pp.25-30 |
SP |
|
Toyama |
Toyama Prefectural University |
Normalization of Vocal Tract Shape Using a Radial Basis Function Jianguo Wei, Xugang Lu, Qiang Fang, Jianwu Dang (JAIST) SP2007-44 |
A simple normalization procedure was applied to Electromagnetic Midsagittal Articulographic (EMMA) data using a radial b... [more] |
SP2007-44 pp.121-126 |