Hello:wave:, I am a 3rd-year Ph.D student in School of Computing at KAIST, advised by Professor Sung-eui Yoon. I received MS from POSTECH and my bachelor’s degree from the university of sheffield.

I love a good story, whether it comes from a film or a book. My research brings this interest to video understanding, from short clips to full-length films and live streams, and from what a scene shows to what a story implies. My current research interests are:

  1. Long-form video understanding: narrative-level reasoning over films and series.
  2. Implicit meaning in video: interpreting metaphor across diverse video formats.
  3. Modeling time in video-language models: positional encoding for long and unbounded video.

News

  • (07/2026) Our paper accepted at ECCV 2026
  • (10/2025) Our team won the 1st place in Grounded VideoQA - Perception Test @ICCV2025
  • (09/2025) I received Ph.D fellowship from NRF
  • (07/2022) Our paper accepted at ECCV 2022.
  • (06/2020) Our team won the 1st place in WSOL - LID challenge @CVPR 2020
  • (07/2019) Our paper accepted at ICCV 2019.

Selected Publications

[3] What You Ask is What You Ground: Bridging Question Intent to Temporal Evidence for Grounded VideoQA.
Jinhwan Seo, Kyubeom Han, Jumin Lee, Junhyug Noh*, Sung-eui Yoon*
ECCV 2026 - paper | project page | code

[2] Object Discovery via Contrastive Learning for Weakly Supervised Object Detection.
Jinhwan Seo, Wonho Bae, Danica J. Sutherland, Junhyug Noh*, Daijin Kim*
ECCV 2022 - paper | project page | code | bibtex

[1] Better to Follow, Follow to be Better: Towards Precise Supervision of Feature Super-Resolution for Small Object Detection.
Junhyug Noh, Wonho Bae, Wonhee Lee, Jinhwan Seo, Gunhee Kim
ICCV 2019 - paper | bibtex

Professional Activities

Conference Reviewer

  • CVPR (2023-current)
  • ICCV/ECCV (2023-current)

Honors & Awards

  • (10/2025) 1st place in Grounded VideoQA - Perception Test @ICCV2025
  • (09/2025) Recipient of Ph.D. Fellowship, NRF
  • (06/2020) 2nd place in WSSS - LID @CVPR2020
  • (06/2020) 1st place in WSOL - LID @CVPR2020