Research Interests

Computer Vision, Image Recognition.

Work Experience

Assistant Professor (Apr. 2023 - present)

Laboratory for Future Interdisciplinary Research of Science and Technology (FIRST), Institute of Integrated Research (IIR), Institute of Science Tokyo (formerly: Tokyo Institute of Technology)
Okumura-Funakoshi Group

Education

Ph.D., (Apr. 2020 - Mar. 2023)

Department of Information and Communication Engineering,

Graduate School of Information Science and Technology,

The University of Tokyo.
Advisor: Prof. Toshihiko Yamasaki

M.S., (Apr. 2018 - Mar. 2020)

Department of Information and Communication Engineering,

Graduate School of Information Science and Technology,

The University of Tokyo.
Advisor: Prof. Toshihiko Yamasaki

B.E., (Apr. 2017 - Mar. 2018)

Department of Information and Communication Engineering,

The University of Tokyo.
Advisor: Prof. Kiyoharu Aizawa and Prof. Toshihiko Yamasaki

Publications

Journals

• Satoshi Kosugi and Toshihiko Yamasaki
  Personalized Image Enhancement Featuring Masked Style Modeling
  IEEE TCSVT, 2024. [pdf] [code]

• Satoshi Kosugi and Toshihiko Yamasaki
  Crowd-Powered Photo Enhancement Featuring an Active Learning Based Local Filter
  IEEE TCSVT, 2023. [pdf] [code]

International Conferences

• Yuanyuan Cai, Satoshi Kosugi, Kotaro Funakoshi, Manabu Okumura
  Enhancing Image Clustering with Captions
  PACLIC, 2024.

• Chun-Fang Chuang, Dongyuan Li, Satoshi Kosugi, Kotaro Funakoshi, Manabu Okumura
  LPLS: A Selection Strategy Based on Pseudo-Labeling Status for Semi-Supervised Active Learning in Text Classification
  PACLIC, 2024.

• Dongyuan Li, Ying Zhang, Zhen Wang, Shiyin Tan, Satoshi Kosugi, Manabu Okumura
  Active Learning for Abstractive Text Summarization via LLM-Determined Curriculum and Certainty Gain Maximization
  EMNLP Findings, 2024.

• Satoshi Kosugi
  Prompt-Guided Image-Adaptive Neural Implicit Lookup Tables for Interpretable Image Enhancement
  ACM Multimedia, 2024. (acceptance rate 26.20%) [pdf] [code]

• Aru Maekawa, Satoshi Kosugi, Kotaro Funakoshi, and Manabu Okumura
  DiLM: Distilling Dataset into Language Model for Text-level Dataset Distillation
  NAACL Findings, 2024.

• Nobukatsu Kajiura, Satoshi Kosugi, Xueting Wang, and Toshihiko Yamasaki
  Self-Play Reinforcement Learning for Fast Image Retargeting
  ACM Multimedia, 2020. (acceptance rate 27.8%)

• Satoshi Kosugi and Toshihiko Yamasaki
  Unpaired Image Enhancement Featuring Reinforcement-Learning-Controlled Image Editing Software
  AAAI, 2020. (acceptance rate 20.6%) [pdf] [code]

• Satoshi Kosugi, Toshihiko Yamasaki, and Kiyoharu Aizawa
  Object-Aware Instance Labeling for Weakly Supervised Object Detection
  ICCV, oral, 2019. (4.3% of all submitted papers or 17% of all accepted papers) [pdf] [code]

Domestic Conferences

• 17 papers

Awards

• IE Award, Technical Committee on Image Engineering, June, 2024
• Committee Special Award, The 30th Annual Meeting of ANLP, 2024 [co-author]
• Yasujiro Niwa Outstanding Paper Award, 2023
• CyberAgent Award, Young researcher Association for NLP Studies (YANS), 2023 [co-author]
• The Best Master Thesis Award, Graduate School of Information Science and Technology, The University of Tokyo, 2019
• MIRU Student Honorable Mention, Meeting on Image Recognition and Understanding (MIRU), 2019
• MIRU Student Honorable Mention, Meeting on Image Recognition and Understanding (MIRU), 2018
• CGVI Best Presentation Award, Technical Committee on Computer Graphics and Visual Informatics, 2018

Invited Talks

• Crowd-Powered Photo Enhancement Featuring an Active Learning Based Local Filter
  The 22nd Forum on Information Technology (FIT2023), 2023.

• Unpaired Image Enhancement Featuring Reinforcement-Learning-Controlled Image Editing Software
  The 23rd Meeting on Image Recognition and Understanding (MIRU), 2020.

• Object-Aware Instance Labeling for Weakly Supervised Object Detection
  Technical Committee on Pattern Recognition and Media Understanding (PRMU), The University of Tokyo, Oct. 2019.

Invited Reviews

• Satoshi Kosugi and Toshihiko Yamasaki
  Object-Aware Instance Labeling for Weakly Supervised Object Detection
  Journal of the Imaging Society of Japan, 2020, 59.6: 585-590.

Funding Sources

• Strategic Basic Research Programs ACT-X, Japan Science and Technology Agency (JST). (Oct. 2024 - Mar. 2027)
• Grant-in-Aid for Research activity start-up, Japan Society for the Promotion of Science (JSPS). (Sep. 2023 - Mar. 2025)
• Research fellowship for young scientists (DC2), Japan Society for the Promotion of Science (JSPS). (Apr. 2021 - Mar. 2023)
• Toyota/Dwango Scholarship for Advanced Artificial Intelligence Researcher. (Apr. 2020 - Mar. 2022)

Links

Okumura-Funakoshi Group

Yamasaki Laboratory

Aizawa Yamakata Matsui Laboratory

---