|
Affiliation |
Faculty of Engineering, Division of Intelligent Systems Engineering |
|
Job Title |
Associate Professor |
|
Research Fields, Keywords |
Communication Science, Multimodal Interaction, Social Signal Processing |
OTSUKA Kazuhiro
|
|
Education 【 display / non-display 】
-
2005.4-2007.3
Nagoya University Doctor Course Completed
-
1993.4-1995.3
Yokohama National University Electrical and Computer Engineering Master Course Completed
-
1989.4-1993.3
Yokohama National University Division of Electrical and Computer Enginnering Graduated
Degree 【 display / non-display 】
-
Doctor of Information Science - Nagoya University
-
Master of Engineering - Yokohama National University
Campus Career 【 display / non-display 】
-
2020.4
Duty Yokohama National UniversityFaculty of Engineering Division of Intelligent Systems Engineering Associate Professor
External Career 【 display / non-display 】
-
2012.10-2020.3
-
2007.4-2012.9
-
2003.4-2007.3
-
2001.4-2003.3
-
1999.12-2001.3
Academic Society Affiliations 【 display / non-display 】
-
2022.4
The Japanese Society for Artificial Intelligence
-
2005
IEEE(Institute of Electrical and Electronics Engineering)
-
1996
電子情報通信学会
-
1996
情報処理学会
Research Areas 【 display / non-display 】
-
Informatics / Intelligent informatics
-
Informatics / Human interface and interaction
Thesis for a degree 【 display / non-display 】
-
時系列画像に基づく動的事象の推測に関する研究
大塚和弘
2007.3
Doctoral Thesis Single Work
Papers 【 display / non-display 】
-
Recognition of Head-Movement Functions in Conversations Using Verbal Features
KIDO Yusei, SUGIYAMA Shunta, OTSUKA Kazuhiro
J107-A ( 1 ) 6 - 18 2024.1 [Reviewed]
Language：Japanese Publishing type：Research paper (scientific journal) Publisher：The Institute of Electronics, Information and Communication Engineers Joint Work
To recognize the functions of head movements in conversations, this paper proposed novel models that integrate conversation participants' verbal and nonverbal behavioral features. As the nonverbal features, these models employ embedding vectors from the CNNs, which have been previously proposed to recognize the head-movement functions. As the verbal features, we examine three representations based on BERT-based embedding vectors in terms of their temporal granularity. Then, a forward neural network integrates these feature vectors to decide the presence or absence of each head-movement function at each time frame in the conversation. The experiments confirmed the proposed model outperformed the previous ones and improved the F-measure by 10.4 points at maximum for a back-channel function.
-
Recognition of Gaze Functions in Conversations from Multimodal Nonverbal Behaviors
TASHIRO Ayane, IMAMURA Mai, KUMANO Shiro, OTSUKA Kazuhiro
J106-A ( 12 ) 296 - 311 2023.12 [Reviewed]
Language：Japanese Publishing type：Research paper (scientific journal) Publisher：The Institute of Electronics, Information and Communication Engineers Joint Work
This paper proposes machine-learning models for recognizing the gaze functions in conversations. For that purpose, we first created a functional gaze corpus with 43 functions, including monitoring, regulation, and expressiveness. Targeting frequent gaze functions, we propose several models using convolutional neural networks (CNNs), which input the multimodal features, including head pose, utterance/silence status, gaze/avert status, eyeball direction, and facial expression. Experimental results confirmed the potentials of the proposed models, which improved the average F-score for 14 functions by 0.243 at maximum and can recognize a gaze function called “looking while listening,” with the F value of 0.839.
-
Deep Regression Models for Estimating Ambiguity in Functional Interpretations of Head Movements in Conversations
TAKEDA Kazuki, IMAMURA Mai, OTSUKA Kazuhiro
J106-A ( 10 ) 255 - 266 2023.6 [Reviewed]
Language：Japanese Publishing type：Research paper (scientific journal) Publisher：The Institute of Electronics, Information and Communication Engineers Joint Work
Head movements play essential roles in human conversations and entail intrinsic ambiguities in expressing and perceiving. This paper proposes regression models for estimating the degree of ambiguities in the head-movement function of conversation participants. As the ambiguities, we introduced expressional perceptibility and interpretational diversity, which are defined as the average and standard deviation of the ratings of multiple observers, respectively. This paper proposes convolutional neural networks, which simultaneously estimate both ambiguities, called μ-σ CNNs. The experiments compared the μ-σ CNNs with other models, which independently estimate each of the ambiguities, and confirmed that the μ-σ CNNs have potential to outperform the other models.
-
Otsuchi Shumpei, Miyoshi Kazuki, Ishii Yoko, Ishiii Ryo, Eitoku Shin-ichiro, Otsuka Kazuhiro
Transactions of the Japanese Society for Artificial Intelligence 38 ( 3 ) H-M74_1 - 13 2023.5 [Reviewed]
Language：Japanese Publishing type：Research paper (scientific journal) Publisher：The Japanese Society for Artificial Intelligence Joint Work
-
Analyzing Synergetic Functions of Listener's Head Movements and Aizuchi in Conversations
Iizuka Kaito, Otsuka Kazuhiro
Transactions of the Japanese Society for Artificial Intelligence 38 ( 3 ) J-M91_1 - 17 2023.5
Language：Japanese Publishing type：Research paper (scientific journal) Publisher：The Japanese Society for Artificial Intelligence Joint Work
Review Papers 【 display / non-display 】
-
非言語情報が生み出すコミュニケーションのリアリティ
大塚和弘
NTT研究所発 触感コンテンツ専門誌ふるえ 8 ( 5 ) 2016.12
Language：Japanese Publishing type：Article, review, commentary, editorial, etc. (other) Single Work
-
Prediction of "Who Will Be Next Speaker and When" in Multi-Party Meetings
Ryo Ishii, Kazuhiro Otsuka, Shiro Kumano, Junji Yamato
NTT Technical Review 13 ( 7 ) 2015
Language：English Publishing type：Internal/External technical report, pre-print, etc. Joint Work
-
The Future of Idobata Kaigi (“Congregation at the Well”): Realizing Natural Conversations with Distance Access
Kazuhiro Otsuka, Nobuhiro Furuyama, Mayumi Bono
NII Today 48 6 - 7 2014.1
Language：English Publishing type：Internal/External technical report, pre-print, etc. Joint Work
-
遠隔地間でも自然な会話を実現する、未来の井戸端会議とは
大塚和弘，古山宣洋，坊農真弓
国立情報学研究所 NII Today 62 6 - 7 2013.12
Language：Japanese Publishing type：Internal/External technical report, pre-print, etc. Joint Work
-
MM-Space: Recreating Multiparty Conversation Space by Using Dynamic Displays
Kazuhiro Otsuka
NTT Technical Review 10 ( 11 ) 2012.11
Language：English Publishing type：Internal/External technical report, pre-print, etc. Single Work
Awards 【 display / non-display 】
-
2022.1 電子情報通信学会 ヒューマンコミュニケーション基礎研究専門委員会 「対話中の非言語行動の機能スペクトラム解析に向けて〜頭部運動を対象とした機能コーパスと深層学習モデルの構築事例〜」
Individual or group name of awards：大塚和弘
-
2021.12 電子情報通信学会 HCGシンポジウム2021 複数人対話における深層転移学習による頭部運動相互作用の機能認識
Individual or group name of awards：森 駿志, 大塚和弘
-
情報処理学会論文誌ジャーナル特選論文賞
2021.1 情報処理学会 話者継続・交替時における対話行為と視線行動に基づく共感スキルの推定
Individual or group name of awards：石井 亮, 大塚和弘, 熊野史朗, 東中竜一郎, 青野裕司
-
電子情報通信学会 HCGシンポジウム2018 特集テーマセッション賞
2018.12 電子情報通信学会
Individual or group name of awards：石井 亮，大塚和弘，熊野史朗，東中竜一郎，富田準二
-
電子情報通信学会 HCGシンポジウム2015 優秀インタラクティブ発表賞
2015.12 電子情報通信学会
Individual or group name of awards：熊野史朗，大塚和弘，石井亮，大和淳司
Presentations 【 display / non-display 】
-
対話中の頭部運動機能の解釈の曖昧性を推定する深層回帰モデルの検討
武田 一輝, 大塚 和弘
第21回情報科学技術フォーラム(FIT2022)
Event date： 2022.9
Language：Japanese Presentation type：Oral presentation (general)
-
複数人対話における深層転移学習による頭部運動相互作用の機能認識
森 駿志, 大塚和弘
電子情報通信学会HCGシンポジウム2021 2021.12 電子情報通信学会
Event date： 2021.12
Language：Japanese Presentation type：Oral presentation (general)
-
Shumepi Otsuchi, Yoko Ishii, Momoko Nakatani, Kazuhiro Otsuka
23rd ACM International Conference on Multimodal Interaction (ICMI2021) 2021.10 ACM
Event date： 2021.10
Language：English Presentation type：Poster presentation
Venue：Montreal, Canada Country：Canada
-
Takashi Mori, Kazuhiro Otsuka
23rd ACM International Conference on Multimodal Interaction (ICMI2021) 2021.10 ACM
Event date： 2021.10
Language：English Presentation type：Poster presentation
Venue：Montreal, Canada Country：Canada
-
Inflation-Deflation Networks for Recognizing Head-Movement Functions in Face-to-Face Conversations
Kazuki Takeda, Kazuhiro Otsuka
23rd ACM International Conference on Multimodal Interaction (ICMI2021) 2021.10 ACM
Event date： 2021.10
Language：English Presentation type：Poster presentation
Venue：Montreal, Canada Country：Canada
Charge of on-campus class subject 【 display / non-display 】
-
2024 Laboratory for Electrical and Computer Engineering
College of Engineering Science
-
2024 Information Theory
College of Engineering Science
-
2024 Innovation and New Business II
Graduate school of Engineering Science
-
2024 Practical System Design
Graduate school of Engineering Science
-
2024 Advanced information system Ⅱ-F
Graduate school of Engineering Science
Committee Memberships 【 display / non-display 】
-
情報処理学会 情報規格調査会
2018.7 - 2020.3 専門委員
Committee type：Academic society
-
電子情報通信学会 ヴァーバル・ノンヴァーバル・コミュニケーション研究会
2016.4 - 2020.3 委員長
Committee type：Academic society
-
電子情報通信学会 和文論文誌D編集委員会
2010.5 - 2014.6 編集委員
Committee type：Academic society
-
電子情報通信学会 ヴァーバル・ノンヴァーバル・コミュニケーション研究会
2010.4 - 2016.3 副委員長（会計）
Committee type：Academic society
-
電子情報通信学会 ヴァーバル・ノンヴァーバル・コミュニケーション研究会
2008.4 専門委員，運営委員
Committee type：Academic society