Takuma Yagi, Ph.D.

• Home
• Projects
• Publications
• Japanese

Takuma Yagi, Ph. D.
Research Scientist, National Institute of Advanced Industrial Science and Technology (AIST)

I am a Research Scientist at Artificial Intelligence Research Center, AIST.
I got my Ph.D. degree at the University of Tokyo in 2022, advised by Prof. Yoichi Sato.
I'm especially interested in developing intelligent systems that perceives and understand our world and using them to make our life better.

（25/04/18）動画像解析に関する技術研修（インターン）生またはリサーチアシスタントを現在募集しています。
ご興味のある方は募集案内をご一読の上メールにてお問い合わせください。

Email: takuma.yagi[at]aist.go.jp
Google Scholar
LinkedIn

Resarch Interests

Computer Vision and Machine Learning for Human Behavior Understanding

• First-Person Vision
• Video Understanding
• Hand-Object Interaction Understanding
• Learning from Human Observation
• Vision & Language
• Assisting Human Activities with Computer Vision

News

• (Nov. 2024) I am visiting Robotics Institute, Carnegie Mellon University as a visiting scholar until next March.
• (Nov. 2024) Our work on object state recognition has been accepted to WACV2025@Tucson! Check out the project page!
• (Feb. 2024) Our work on a video dataset of biological experiments has been released! Check out the preprint and dataset.
• (Apr. 2023) I started working as a research scientist at AIST!

Selected Publications

New!
Learning Multiple Object States from Actions via Large Language Models
Masatoshi Tateno, Takuma Yagi, Ryosuke Furuta, and Yoichi Sato
WACV2025, LPVL Workshop@CVPR2024
[project page] [preprint]

FineBio: A Fine-Grained Video Dataset of Biological Experiments with Hierarchical Annotation
Takuma Yagi, Misaki Ohashi, Yifei Huang, Ryosuke Furuta, Shungo Adachi, Toutai Mitsuyama, and Yoichi Sato
EPIC Workshop@CVPR2023
[preprint][data & code]

Ego-Exo4D: Understanding Skilled Human Activity from First- and Third-Person Perspectives.
Kristen Grauman, Andrew Westbury, Lorenzo Torresani, Kris Kitani, Jitendra Malik, ..., Takuma Yagi, ..., Michael Wray
CVPR 2024
[project page] [preprint]

Fine-grained Affordance Annotation for Egocentric Hand-Object Interaction Videos
Zecheng Yu, Yifei Huang, Ryosuke Furuta, Takuma Yagi, Yusuke Goutsu, and Yoichi Sato.
WACV 2023, EPIC Workshop@CVPR2022
[preprint][proceedings]

Ego4D: Around the World in 3,000 Hours of Egocentric Video
Kristen Grauman, Andrew Westbury, Eugene Byrne, Zachary Chavis, Antonino Furnari, Rohit Girdhar, ..., Takuma Yagi, ... Jitendra Malik.
CVPR 2022 (Oral, Best Paper Finalist)
[project page] [paper]

Hand-Object Contact Prediction via Motion-Based Pseudo-Labeling and Guided Progressive Label Correction
Takuma Yagi, Md. Tasnimul Hasan, and Yoichi Sato.
BMVC 2021
[paper & talk] [preprint] [data & code]

GO-Finder: Registration-Free Wearable System for Assisting Users in Finding Lost Objects via Hand-Held Object Discovery
Takuma Yagi, Takumi Nishiyasu, Kunimasa Kawasaki, Moe Matsuki, and Yoichi Sato.
IUI 2021 (Long Paper), TiiS'22
[Preprint] [Paper] [Video] [Presentation]

Future Person Localization in First-Person Videos
Takuma Yagi, Karttikeya Mangalam, Ryo Yonetani, and Yoichi Sato.
CVPR 2018 (Spotlight)
[Preprint] [Paper] [Data & Code]

Work Experience

Nov. 2024 - Mar. 2025, Visitng Scholar, Robotics Institute, Carnegie Mellon University

Research on assembly action understanding from videos (Mentor: Prof. Kris Kitani)

Apr. 2023 - Present, Research Scientist, Artificial Intelligence Research Center, National Institute of Advanced Industrial Science and Technology (AIST)

Research Scientist at Computer Vision Research Team

Apr. 2023 - Present, Cooperative Research Fellow, Institute of Industrial Science, The University of Tokyo

Cooperative Research Fellow at Sato Laboratory

Apr. 2022 - Mar. 2023, Project Researcher, Institute of Industrial Science, The University of Tokyo

Analysis of biochemical experiment videos at Sato Laboratory

Oct. 2022 - Mar. 2023, Visiting Researcher, Nihon University

Research on Human-Agent Interaction at Osawa Laboratory

Sep. 2018 - Dec. 2018, Microsoft Research Asia

Research Internship at Media Computing Group. Advisor: Dr. Xun Guo

Jun. 2017 - Mar. 2018, Dwango Artificial Intelligence Laboratory

Curriculum supervisor/lecturer at Neuro-Inspired COmputing To AI School (NICO2AI School)

Dec. 2015 - Jan. 2016, Datasection.Inc

Machine Learning Engineer

Mar. 2014 - Mar. 2015, Datasection.Inc

Java/Javascript Engineer

Education

2022 Ph.D. in Information Science and Technology, The University of Tokyo

  Advisor: Prof. Yoichi Sato

  Thesis: Hand-Object Interaction Mining from First-Person Videos

2019 M.S. in Information and Science and Technology, The University of Tokyo

  Advisor: Prof. Yoichi Sato

  Thesis: Future Person Localization in First-Person Videos

2017 B.E. in Computer Science, Tokyo Institute of Technology

  Advisor: Prof. Koichi Shinoda

  Thesis: Patch-based Human Action Recognition from Video via Disrciminative Feature Mining

Awards

Aug. 2021 MIRU Presentation Award

Mar. 2019 Department Chair's Award for Excellent Master Thesis, Department of Information and Communication Engineering, the University of Tokyo

Grants

Oct. 2022 - Mar. 2025 JST ACT-X AI powered Research Innovation / Creation, 6,000,000 JPY

Oct. 2022 - Mar. 2024 JSPS Grant-in-Aid for Research Activity Start-up, 2,200,000 JPY

Apr. 2021 - Mar. 2022 JSPS Research Fellowship for Young Scientists (JSPS DC2), 800,000 JPY

2019 JASSO Scholarship for Top 10% Excellent Master Students from Japan Student Services Organization, 2,112,000 JPY

Jul. 2018 - Jun. 2023 Masason Foundation Full Member

2018 The NEC C&C Foundations: Grants for Attendees at International Conference, 200,000 JPY

Apr. 2017 - Mar. 2018 Toyota/Dwango Scholarship for Advanced Artificial Intelligence Researcher, 1,200,000 JPY

Services

Reviewer

Organizer

First Joint Egocentric Vision (EgoVis) Workshop Held in Conjunction with CVPR 2024

Links

• GitHub
• Facebook

Last Update: 2025-04-18
© 2025 Takuma Yagi