Hojun Jang
I’m a Ph.D. student at Seoul National University, fortunate to be under the guidance of Professor Young Min Kim. My research focuses on 3D human motion — spanning reconstruction from sparse or noisy sensor inputs (e.g., IMUs, egocentric cameras, partial point clouds) and generation from diverse conditioning signals (e.g., text, audio, or target keyframes). I am particularly interested in learning general-purpose motion representations that transfer across tasks such as pose estimation, motion generation, and motion inbetweening.
News
May. 2026
- I am joining Naver Labs Europe 3D Humans Team as PhD Intern.
- ScaleMoGen and EgoForce are released on arXiv.
Apr. 2025
- Audio-aided Character Control is accepted to Eurographics 2025 short paper.
Sep. 2024
- ReMP is accepted to WACV 2025.