Seoha Kim

Post-Master's Research Fellow

kim-seoha@naver.com

About

I am a post-master’s research fellow interested in 3D/4D scene understanding, scene graphs, generalizable reconstruction, and robotics.
Currently, my research focuses on generalizable dynamic 3D Gaussian Splatting using egocentric videos.
My long-term goal is to develop an intelligent system that can perceive and interact with dynamic 3D environments.

Coffee chats about research are always welcome! ☕️

Publications

indicates equal contribution.

Efficient 4D Scaffold Gaussian Splatting with Dynamic‑Aware Anchor Growing

Woong Oh Cho, In Cho, Seoha Kim, Jeongmin Bae, Youngjung Uh, Seon Joo Kim

Under Review

Rethinking Open-Vocabulary Segmentation of Radiance Fields in 3D Space

Hyunjee Lee, Youngsik Yun, Jeongmin Bae, Seoha Kim, Youngjung Uh

AAAI 2025

Per-Gaussian Embedding based Deformation for Deformable 3D Gaussian Splatting

Jeongmin Bae, Seoha Kim, Youngsik Yun, Hahyun Lee, Gun Bang, Youngjung Uh

ECCV 2024

Optimizing Dynamic NeRF and 3DGS with No Video Synchronization

Seoha Kim, Jeongmin Bae, Youngsik Yun, Hahyun Lee, Gun Bang, Youngjung Uh

ECCV 2024 Wild3D Workshop

Sync-NeRF: Generalizing Dynamic NeRFs to Unsynchronized Videos

Seoha Kim, Jeongmin Bae, Youngsik Yun, Hahyun Lee, Gun Bang, Youngjung Uh

AAAI 2024

Vitae

  • ETRI 2024.02 - 2025.1
    Post-Master's Research Fellow
    Dynamic Scene Reconstruction
  • ETRI 2023.1 - 2024.6
    Academic-Research Cooperation
    Dynamic Scene Reconstruction Research
  • LG Display 2022.2 ‑ 2022.12
    Academic-Industrial Cooperation
    Knowledge Distillation Research
  • Yonsei University 2022 – 2024
    M.S. Student
    Artificial Intelligence
  • Plask 2021.3 – 2021.8
    AI Engineer
    3D Pose Estimation Research
  • Hyundai Mobis 2019.9 – 2020.2
    Data Scientist, Intern
    Internship in Data Science Team
  • Yonsei University 2015 - 2021
    B.A. Student
    Business Administration

Patents

KR 10-2024-0043684, Method and apparatus for Dynamic Gaussian Splatting using embedding-based deformation
KR 10-2023-0105173, Apparatus for representing dynamic neural radiance fields from unsynchronized videos
KR 10-2020-0022362, Apparatus of diagnosing noise quality of motor