ken-system: Advance Program - 2023-02-SP-IPSJ-SLP-EA-SIP

IEICE Technical Committee Submission System
Advance Program

Online Proceedings
[Sign in]
Tech. Rep. Archives

===============================================
Technical Committee on Engineering Acoustics (EA)
Chair: Kenichi Furuya (Oita Univ.) Vice Chair: Tatsuya Kako (NTT), Junki Ono (Tokyo Metropolitan Univ.)
Secretary: Keigo Wakayama (NTT), Takanobu Nishiura (RitsumeikanUniv.)
Assistant: Masato Nakayama (OSU), Kouhei Yatabe (Tuat)

===============================================
Technical Committee on Signal Processing (SIP)
Chair: Toshihisa Tanaka (Tokyo Univ. Agri.&Tech.)
Vice Chair: Koichi Ichige (Yokohama National Univ.), Takayuki Nakachi (Ryukyu Univ.)
Secretary: Yuichi Tanaka (Tokyo Univ. Agri.&Tech.), Seisuke Kyochi (Univ. of Kitakyushu)
Assistant: Taichi Yoshida (UEC), Shoko Imaizumi (Chiba Univ.)

===============================================
Technical Committee on Speech (SP)
Chair: Tomoki Toda (Nagoya Univ.)
Secretary: Ryo Masumura (NTT), Toru Nakashika (Univ. of Electro-Comm.)
Assistant: Ryo Aihara (Mitsubishi Electric), Daisuke Saito (Univ. of Tokyo)

===============================================
Special Interest Group on Spoken Language Processing (IPSJ-SLP)
Chair: Tomoki Toda (Nagoya Univ.)
Secretary: Ryo Masumura (NTT), Toru Nakashika (Univ. of Electro-Comm.)
Assistant: Ryo Aihara (Mitsubishi Electric), Daisuke Saito (Univ. of Tokyo)

DATE:
Tue, Feb 28, 2023 09:10 - 17:35
Wed, Mar 1, 2023 09:10 - 17:40

PLACE:
Okinawa Prefectural Museum & Art Museum(3-1-1 Omoromachi, Naha-shi, Okinawa 900-0006. https://okimu.jp/guide/access/. Prof. Takayuki Nakachi. 098-941-8200)

TOPICS:

----------------------------------------
Tue, Feb 28 AM SP1 (09:10 - 10:30)
----------------------------------------

(1)/SP 09:10 - 09:30
Comparison of fundamental frequency controllable fast neural waveform generative models.
Sota Shimizu (Kobe Univ./NICT), Takuma Okamoto (NICT), Ryoichi Takashima, Tetsuya Takiguchi (Kobe Univ.), Tomoki Toda (Nagoya Univ./NICT), Hisashi Kawai (NICT)

(2)/SP 09:30 - 09:50
MS-FC-HiFiGAN : Fast Neural Waveform Generation Model With Learnable Lightweight Upsampling
Haruki Yamashita (Kobe Univ/NICT), Takuma Okamoto (NICT), Ryoichi Takashima, Tetsuya Takiguchi (Kobe Univ), Tomoki Toda (Nagoya Univ/NICT), Hisashi Kawai (NICT)

(3)/SP 09:50 - 10:10
End-to-End Speech Synthesis Based on Articulatory Movements Captured by Real-time MRI
Yuto Otani, Shun Sawada, Hidefumi Ohmura, Kouichi Katsurada (Tokyo Univ. Sci.)

(4)/SP 10:10 - 10:30
Singing voice synthesis based on a frame-driven attention mechanism considering vocal timing deviation
Miku Nishihara, Yukiya Hono, Kei Hashimoto, Yoshihiko Nankaku, Keiichi Tokuda (NITech)

----------------------------------------
Tue, Feb 28 AM EA1 (09:10 - 10:30)
----------------------------------------

(5)/EA 09:10 - 09:30
Extension of acoustic system measurement based on signal safeguarding
-- Repetition and orthogonalization for post hoc analysis --
Hideki Kawahara (和歌山大), Kohei Yatabe (Tokyo Univ。 of Agriculture and Technology), Ken-Ichi Sakakibara (Health Sciences Univ. of Hokkaido), Mitsunori Mizumachi (Kyushu Inst. of Tech.)

(6)/EA 09:30 - 09:50
A Study on Designing Hopping Patterns Based on Euler Graphs for Inaudible Sound Communication Systems
Naofumi Aoki, Kosei Ozeki (Hokkaido Univ.), Kenichi Ikeda, Hiroshi Yasuda, Hiroyuki Namba (SST)

(7)/EA 09:50 - 10:10
Influence of Reflections in Small-Scale Anechoic Room Measurements
Tatsuya Higuchi, Yutaka Kaneda, Kenji Suyama (Tokyo Denki Univ.)

(8)/EA 10:10 - 10:30
Generation of the individualized head-related transfer functions in the upper hemisphere using parametric notch-peak model in the median plane
Fuka Nakamura, Kazuhiro Iida (CIT)

----- Break ( 10 min. ) -----

----------------------------------------
Tue, Feb 28 AM SLP (10:40 - 12:00)
----------------------------------------

(9) 10:40 - 11:00
SLP

(10) 11:00 - 11:20
SLP

(11) 11:20 - 11:40
SLP

(12) 11:40 - 12:00
SLP

----------------------------------------
Tue, Feb 28 AM SIP1 (10:40 - 12:00)
----------------------------------------

(13)/SIP 10:40 - 11:00
Image reconstruction with a diffusion model for robust image classification against unknown degradation
Teruaki Akazawa (Tokyo Metro. Univ.), Yuma Kinoshita (Tokai Univ.), Hitoshi Kiya (Tokyo Metro. Univ.)

(14)/SIP 11:00 - 11:20
The target detection method through autocovariance matrices and its robust analysis
Yusuke Ono, Linyu Peng (Keio Univ.)

(15)/SIP 11:20 - 11:40
Hadamard-coded Supervised Discrete Hashing on Quaternion Domain
Akari Katsuma, Seisuke Kyochi (Kogakuin Univ.), Shunsuke Ono (Tokyo Tech.), Ivan Selesnick (New York Univ.)

(16)/SIP 11:40 - 12:00
Acoustic Echo and Noise Canceller Based on Minimization of Shared-Error Signal
Kenta Iwai, Takanobu Nishiura (Ritsumeikan Univ.)

----- Lunch Break ( 60 min. ) -----

----------------------------------------
Tue, Feb 28 PM Invited Talk 1 (13:00 - 13:45)
----------------------------------------

(17)/SP 13:00 - 13:45
[Invited Talk]
Multiple sound spot synthesis meets multilingual speech synthesis
-- Implementation is really all we need --
Takuma Okamoto (NICT)

----------------------------------------
Tue, Feb 28 PM Invited Talk 2 (13:45 - 14:30)
----------------------------------------

(18)/EA 13:45 - 14:30
[Invited Talk]
Multichannel audio source separation based on deep generative model and signal independence
Li Li (CA)

----- Break ( 10 min. ) -----

----------------------------------------
Tue, Feb 28 PM Short Presentation 1 (14:40 - 15:45)
----------------------------------------

(19) 14:40 - 14:45
Design of a regularization of blind deblurring for blurred images containing saturated pixels and Gaussian noise
Tomoya Kobayashi, Ryo Hayakawa, Youji Iiguni (Osaka Univ.)

(20) 14:45 - 14:50
Application of Deep Unfolding to Video Reconstruction Algorithms from Compressed Images
Takashi Matsuda, Ryo Hayakawa, Youji Iiguni (Osaka Univ.)

(21) 14:50 - 14:55
Consideration of misalignment in multi-focus image fusion using convolutional sparse representation
Ryo Tamaki, Ryo Hayakawa, Youji Iiguni (Osaka Univ.)

(22) 14:55 - 15:00
Pop Noise Based Speaker Verification with Continuous Phoneme-Pop Data and GBDT
Kenta Takemae, Ryota Shimokura, Yoji Iiguni (OU)

(23) 15:00 - 15:05
A Study on Regularization in Video Super-Resolution Based on LMS Algorithm
Ryogo Shimizu, Ryo Hayakawa, Youji Iiguni (Osaka Univ.)

(24) 15:05 - 15:10
Identification of Seizure Onset Zone from Intracranial EEG Using Source Selection-Based Domain Adaptation
Keisuke Matsubayashi (TUAT), Yasushi Iimura, Takumi Mitsuhashi, Hidenori Sugano (Juntendo Univ.), Kosuke Fukumori, Toshihisa Tanaka (TUAT)

(25) 15:10 - 15:15
Speech synthesis from electrocorticogram using pre-trained neural vocoder
Kai Shigemi, Shuji Komeiji (TUAT), Takumi Mitsuhashi, Yasushi Iimura, Hiroharu Suzuki, Hidenori Sugano (Juntendo Univ.), Koichi Shinoda (Tokyo Tech), Kohei Yatabe, Toshihisa Tanaka (TUAT)

(26) 15:15 - 15:20
Effects of movement on the EEG during rhythmic response
Hiroki Arai, Ingon Chanpornpakdi, Toshihisa Tanaka (TUAT)

(27) 15:20 - 15:25
Single-channel environmental sound classification using distance-based sound separation
Ryoya Ogura, Sayaka Shiota (Tokyo Metropolitan Univ.), Keisuke Imoto (Doshisha Univ.), Hitoshi Kiya (Tokyo Metropolitan Univ.)

(28) 15:25 - 15:30
Comfortable sound design of dental treatment sound based on automatic chord progression generation with modulation conditions using critical bandwidth
Takuya Hayashi, Toru Takahashi, Masato Nakayama (Osaka Sangyo Univ.)

(29) 15:30 - 15:35
Fine-tuning for Speaker Diarization :Measuring Accuracy in Japanese Conversation
Yurina Machida (Tsukuba Univ.), Taishi Yamaoka (Empath)

(30) 15:35 - 15:40
Any-to-Many Voice Conversion with Voice Similarity Comparison and Many-to-Many Model
Hiroaki Hyodo, Tetsuya Sakai (Waseda Univ.)

(31) 15:40 - 15:45
Cross-language Speaker Recognition for Japanese-English Bilinguals
Ryotaro Sano (Chiba Univ.), Masahumi Nishida (Shizuoka Univ), Satoru Tsuge (Daido Univ.), Shingo Kuroiwa, Hiroyuki Yoshimura (Chiba Univ.)

----- Break ( 10 min. ) -----

----------------------------------------
Tue, Feb 28 PM SP-EA (15:55 - 17:35)
----------------------------------------

(32)/SP 15:55 - 16:15
Self-Supervised Learning With Spatial Audio-Visual Recording for Sound Event Localization and Detection
Yoto Fujita (Kyoto Univ.), Yoshiaki Bando (AIST), Keisuke Imoto (Doshisha Univ./AIST), Masaki Onihsi (AIST), Yoshii Kazuyoshi (Kyoto Univ.)

(33)/SP 16:15 - 16:35
Visual onoma-to-wave: environmental sound synthesis from visual onomatopoeias and sound-source images
Hien Ohnaka (NITTC), Shinnosuke Takamichi (UT), Keisuke Imoto (DU), Yuki Okamoto (Rits), Kazuki Fujii, Hiroshi Saruwatari (UT)

(34)/SP 16:35 - 16:55
Generalized warping based on Lie group theory
Atsushi Miyashita, Tomoki Toda (Nagoya Univ.)

(35)/SP 16:55 - 17:15
Vocal tract length estimation using fundamental frequency adaptive auditory representation
Toshio Irino, Shintaro Doan (Wakayama Univ.)

(36)/EA 17:15 - 17:35
DNN-based Noise Reduction Using Noise Signal for Target Signal
Ryota Hiromasa, Hien Ohnaka, Ryoichi Miyazaki (NITTC)

----------------------------------------
Tue, Feb 28 PM EA-SIP (15:55 - 17:35)
----------------------------------------

(37)/EA 15:55 - 16:15
A new configuration of 1-2-2 multi-channel active noise control system
Kensaku Fujii (Kodaway Lab.), Mitsuji Muneyasu (Kansai Univ.), Yoshifumi Chisaki (CIT)

(38)/EA 16:15 - 16:35
A method of constantly estimating the feedback path in active noise control systems
Kensaku Fujii (kodaway Lab.), Mitsuji Muneyasu (Kansai Univ.), Yoshifumi Chisaki (CIT)

(39)/SIP 16:35 - 16:55
Sound Source Localization Method based on Suppression Amount of Complex Weighted Sum Circuit
Tsukasa Hidaka, Kenji Suyama (Tokyo Denki Univ.)

(40)/SIP 16:55 - 17:15
Application of Frequency Domain Adaptive Filter to Residual Noise Reduction
Kai Furusawa, Kenji Suyama (Tokyo Denki Univ.)

(41)/SIP 17:15 - 17:35
A Study of the Number of Groups for CSD Coefficient FIR Filter Design by Grouped ACO
Marika Morikawa, Kenji Suyama (Tokyo Denki Univ.)

----------------------------------------
Wed, Mar 1 AM SP2 (09:10 - 10:30)
----------------------------------------

(42)/SP 09:10 - 09:30
Training Dialect Speech Recognition Model using Corpus of Japanese Dialects and Self-Supervised Learning-based Model XLSR
Shogo Miwa, Atsuhiko Kai (Shizuoka Univ.)

(43)/SP 09:30 - 09:50
A Study on Scheduled Sampling for Neural Transducer-based ASR
Takafumi Moriya, Takanori Ashihara, Hiroshi Sato, Kohei Matsuura, Tomohiro Tanaka, Ryo Masumura (NTT)

(44)/SP 09:50 - 10:10
Domain Adaptation for Improving End-to-end ASR Performance of Classroom Speech with Variable Recording Condition
Raufun Nahar, Rino Suzuki, Atsuhiko Kai (Shizuoka Univ.)

(45)/SP 10:10 - 10:30
Vocabulary-Set Decomposition and Multi-task Learning for Target Vocabulary Extraction in Japanese Speech Recognition
Aoi Ito (LINE/Hosei Univ.), Tatsuya Komatsu, Yusuke Fujita (LINE)

----------------------------------------
Wed, Mar 1 AM EA2 (09:10 - 10:30)
----------------------------------------

(46)/EA 09:10 - 09:30
Joint analysis of acoustic scenes and sound events based on semi-supervised learning
Ami Igarashi, Shunsuke Tsubaki, Keisuke Imoto (DU)

(47)/EA 09:30 - 09:50
Texture Reproduction of Ultrasonic Mid-Air Haptics Based on Amplitude Modulation Signal Generation Using Fricative Sounds Feature Extraction and Hand Tracking
Asuto Ueda, Toru Takahashi, Masato Nakayama (Osaka Sangyo Univ.)

(48)/EA 09:50 - 10:10
Regularization Term Design Based on Spectrogram Consistency in Independent Low-Rank Matrix Analysis for Multichannel Audio Source Separation
Sota Misawa, Norihiro Takamune (UTokyo), Kohei Yatabe (TUAT), Daichi Kitamura (NIT, Kagawa), Hiroshi Saruwatari (UTokyo)

(49)/EA 10:10 - 10:30
Anomalous sound detection with complex-valued hybrid neural networks considering phase variations
Shota Nishiyama, Akira Tamamori (AIT)

----- Break ( 10 min. ) -----

----------------------------------------
Wed, Mar 1 AM SP3 (10:40 - 12:00)
----------------------------------------

(50)/SP 10:40 - 11:00
Diffusion-based parallel voice conversion with source-feature condition
Takuya Kishida, Toru Nakashika (UEC)

(51)/SP 11:00 - 11:20
Representation and Prediction of Accent Phrase Prosodic Features in Japanese Text-to-Speech
Masaki Sato, Shinnosuke Takamichi, Hiroshi Saruwatari (The Univ. of Tokyo)

(52)/SP 11:20 - 11:40
An Investigation of Text-to-Speech Synthesis Using Voice Conversion and x-vector Embedding Sympathizing Emotion of Input Audio for Spoken Dialogue Systems
Shunichi Kohara, Masanobu Abe, Sunao Hara (Okayama Univ.)

(53)/SP 11:40 - 12:00
Choral Singing Voice Synthesis with Modulation Acoustic Features
Sora Miyazawa, Anan Kikuchi, Daisuke Saito, Nobuaki Minematsu (UTokyo)

----------------------------------------
Wed, Mar 1 AM EA3 (10:40 - 12:00)
----------------------------------------

(54)/EA 10:40 - 11:00
Quasi-real-time estimation of a maximum radiation direction from a loudspeaker surrounded by four microphones based on SPL ratio
Ryusei Tsuda, Daiki Maekawa, Tomoru Awatani, Masato Nakayama, Toru Takahashi (Osaka Sangyo Univ.)

(55)/EA 11:00 - 11:20
Analysis of Noisy-target Training for DNN-based speech enhancement and investigation towards its practical use
Takuya Fujimura, Tomoki Toda (Nagoya Univ.)

(56)/EA 11:20 - 11:40
A Study on Selective Fixed-Filter ANC Using 2D-CNN with Sliding DCT input
Kenya Doi, Yoshinobu Kajikawa (KU)

(57)/EA 11:40 - 12:00
Predominant Instrument Recognition in Polyphonic Music Based on Transfer Learning with Vanilla ResNet-50
Lifan Zhong, Daisuke Saito, Nobuaki Minematsu (UTokyo)

----- Lunch Break ( 60 min. ) -----

----------------------------------------
Wed, Mar 1 PM Invited Talk 3 (13:00 - 13:45)
----------------------------------------

(58)/SP 13:00 - 13:45
[Invited Talk]
What Do Self-Supervised Speech Representation Models Know?
-- A Layer-Wise Analysis --
Karen Livescu, Ankita Pasad, Ju-Chieh Chou, Bowen Shi (TTI-Chicago)

----------------------------------------
Wed, Mar 1 PM Invited Talk 4 (13:45 - 14:30)
----------------------------------------

(59)/SP 13:45 - 14:30
[Invited Talk]
Speech and Language Research in the Google Tokyo Office
Michiel Bacchiani (Google)

----- Break ( 10 min. ) -----

----------------------------------------
Wed, Mar 1 PM Short Presentation 2 (14:40 - 15:40)
----------------------------------------

(60) 14:40 - 14:45
Anomalous sound detection based on differential features of multi channel acoustic signals considering spatial and temporal variations
Shota Nishiyama, Akira Tamamori (AIT)

(61) 14:45 - 14:50
Personality Recognition on Dyadic Interactions with Representation Learning
Nathania Nah (Tokyo Tech), Takafumi Koshinaka (YCU), Koichi Shinoda (Tokyo Tech)

(62) 14:50 - 14:55
Corpus construction toward multi-domain empathetic dialogue speech synthesis
Yuki Saito, Eiji Iimori, Shinnosuke Takamichi (UT), Kentaro Tachibana (LINE), Hiroshi Saruwatari (UT)

(63) 14:55 - 15:00
A faster method for blind source separation based on frequency bin selection and linear interpolation
Yuki Nakamura, Ryoichi Miyazaki (NITTC)

(64) 15:00 - 15:05
Self-localization of microphone array in distributed microphone arrays and real environmental experiment using Blinky
Manami Nakamura, Ryoichi Miyazaki (NITTC)

(65) 15:05 - 15:10
Construction of Language Model for Low-resource Domain Speech Recognition Based on Sentence Generation
Ryo Maejima, Daiki Mori, Youkoh Wakabayashi, Norihide Kitaoka (TUT)

(66) 15:10 - 15:15
Automatic Speech Recognition model using data with verbal and non-verbal information tag
Nagito Shione, Yukoh Wakabayashi, Norihide Kitaoka (TUT)

(67) 15:15 - 15:20
Directivity Control of Multichannel One-point Spherical Microphone by Long Short-term Memory Networks
Shota Naiki, Kenta Iwai, Takanobu Nishiura (Ritsumeikan Univ), Yoshiharu Soeta (AIST)

(68) 15:20 - 15:25
Study of Frequency Response Analysis of Effect Cymbals by Finite Element Method
Kohei Izawa, Yuting Geng, Kenta Iwai, Takanobu Nishiura (Ritsumeikan Univ.)

(69) 15:25 - 15:30
Study of Speech Quality Improvement for Interpolated Missing Segments of Extracted Speech Signals from Captured Videos with Dual Rolling-Shutter Cameras
Hayata Nakano, Yuting Geng, Kenta Iwai, Takanobu Nishiura (Ritsumeikan Univ.)

(70) 15:30 - 15:35
(See Japanese page.)

(71) 15:35 - 15:40
Development of chemical terminology learning materials for learners with foreign roots
Mayu Tokumoto, Akemi Ishii (SIT)

----- Break ( 10 min. ) -----

----------------------------------------
Wed, Mar 1 PM SP4 (15:50 - 17:30)
----------------------------------------

(72)/SP 15:50 - 16:10
The linguistic influence on speaker verification based on Self-Supervised Learning
Tomoka Wakamatsu (Tokyo Metropolitan Univ.), Atsushi Ando (NTT), Sayaka Shiota (Tokyo Metropolitan Univ.), Ryo Masumura (NTT), Hitoshi Kiya (Tokyo Metropolitan Univ.)

(73)/SP 16:10 - 16:30
Increasing speech intelligibility for evacuation guidance by mimicking professional announcers' voice
-- Discussion on speech intelligibility and its physical correlates --
KimDung Tran, Masato Akagi, Masashi Unoki (JAIST)

(74)/SP 16:30 - 16:50
Data cleansing using synthetic speech detection for speaker verification
Kenzo Wada, Sayaka Shiota, Hitoshi Kiya (Tokyo Metropolitan Univ.)

(75)/SP 16:50 - 17:10
Effects of Voice Artificiality on the Degree of Compatibility between Voice and Appearance of Voice Agents
Kota Iura, Naotake Masuda, Daisuke Saito, Nobuaki Minematsu (UTokyo)

(76)/SP 17:10 - 17:30
Quantification of Voice Register Information including Mixed Voice based on Class Posterior Probabilities
Yu Kitamura, Anan Kikuchi, Daisuke Saito, Nobuaki Minematsu (UTokyo)

----------------------------------------
Wed, Mar 1 PM SIP2 (15:50 - 17:40)
----------------------------------------

(77)/SIP 15:50 - 16:10
Multiscale Manifold Clustering and Embedding with Multiple Kernels
Kyohei Suzuki, Masahiro Yukawa (Keio Univ.)

(78)/SIP 16:10 - 16:30
On Design of Real Filters For Directed Graph Signals
Shogo Muramatsu, Hotaka Kitamura, Hiroyashu Yasuda (Niigta Univ.), Yuichi Tanaka (Osaka Univ.)

(79)/SIP 16:30 - 16:50
Low-bit Image Restoration with Loop-unrolled ISTA
Shu Abe, Soushi Takahashi, Shogo Muramatsu (Niigata Univ)

(80)/SIP 16:50 - 17:10
A Study on Virtual Sensing Method for Hybrid Active Noise Control System
Shota Toyooka, Kajikawa Yoshinobu (Kansai Univ.)

(81)/SIP 17:10 - 17:30
RGB-D Salient Object Detection Using Saliency and Edge Reverse Attention
Tomoki Ikeda, Masaaki Ikehara (Keio Univ.)

(82) 17:30 - 17:40
Closing

# Information for speakers
General Talk will have 15 minutes for presentation and 5 minutes for discussion.

# CONFERENCE SPONSORS:
- This conference is co-sponsored by Technical Committee on Electroacoustics of ASJ and APSIPA Japan Chapter. This conference is technical co-sponsored by IEEE SPS Tokyo Joint Chapter.

=== Technical Committee on Engineering Acoustics (EA) ===
# FUTURE SCHEDULE:

Thu, May 25, 2023: Online [Wed, Mar 15]

# SECRETARY:
Keigo Wakayama (NTT)
E-mail:

hco

=== Technical Committee on Signal Processing (SIP) ===
# FUTURE SCHEDULE:

Thu, May 18, 2023 - Fri, May 19, 2023: Sansui Hall, Mie University [Thu, Mar 23]

# SECRETARY:
IEICE Technical Group on Signal Processing
Email: sip-