Yu-Ji Kim

I am a Ph.D. student at POSTECH.
I have worked on embodied reasoning and exploration based on spatial understanding. My research interests lie in enabling embodied agents to perceive and act in the physical world to assist humans.

Email / CV / Scholar / LinkedIn / GitHub

News

2026/06: One paper has been accepted to ECCV 2026.
2026/02: I'll start a research internship at NVIDIA Research Taiwan.
2025/10: One have been selected as an Outstanding Reviewer (Top 3%) at ICCV 2025.
2025/03: One paper has been accepted to CVPR 2025.
2024/11: Our paper MeTTA won the Best Poster Award at BMVC 2024.
2024/07: One paper has been accpted to BVMC 2024.
2023/07: One paper has been accepted to ICCV 2023.
2022/07: One paper has been accepted to ECCV 2022.

Research

	Enhancing Embodied Reasoning and Grounding by Novel View Synthesis Kim Yu-Ji, Dahye Lee, Kim Jun-Seong, Nam Hyeon-Woo, GeonU Kim, Yongjin Kwon, Yu-Chiang Frank Wang, Jaesung Choe, Tae-Hyun Oh ECCV, 2026 Embodied reasoning and grounding by integrating 3DGS and VLMs.
	Dr. Splat: Directly Referring 3D Gaussian Splatting via Direct Language Embedding Registration Kim Jun-Seong, GeonU Kim, Kim Yu-Ji, Yu-Chiang Frank Wang, Jaesung Choe^†, Tae-Hyun Oh^† CVPR, 2025 Project Page / Video / arXiv - Accepted as a highlight paper - Winner of Qualcomm Innovation Fellowship 2025 Korea Direct 3D-aware feature embedding with 3DGS.
	MeTTA: Single-View to 3D Textured Mesh Reconstruction with Test-Time Adaptation Kim Yu-Ji, Hyunwoo Ha, Kim Youwang, Jaeheung Surh, Hyowon Ha^†, Tae-Hyun Oh^† BMVC, 2024 Project Page / Video / arXiv - Best Poster Award at BMVC 2024 Single view to 3D object reconstruction with realistic PBR textures.
	Scratching Visual Transformer's Back with Uniform Attention Nam Hyeon-Woo, Kim Yu-Ji, Byeongho Heo, Dongyoon Han, Seong Joon Oh, Tae-Hyun Oh ICCV, 2023 Project Page / Video / arXiv Infusing dense attention splits the responsibility of interactions; the burden of interactions of self-attention is reduced.
	HDR-Plenoxels: Self-Calibrating High Dynamic Range Radiance Fields Kim Jun-Seong, Kim Yu-Ji*, Moon Ye-Bin, Tae-Hyun Oh ECCV*, 2022 Project Page / Video / arXiv High dynamic range (HDR) radiance fields that learn a plenoptic function of 3D HDR radiance fields, geometry information, and varying camera settings inherent in 2D low dynamic range (LDR) images.

Experience

NVIDIA, Taipei, Taiwan

Feb 2026 - Present

Research scientist internship (Manager: Jaesung Choe)
POSTECH, Pohang, Korea

Sep 2021 - Present

Ph.D. candidate (Advisor: Tae-Hyun Oh)

Honors & Awards

Outstanding Reviewer, ICCV 2025
Best Poster Award, BMVC 2024

Invited Talks

HDR-Plenoxels: High Dynamic Range Radiance Fields, Google ExploreCSR Workshop on POSTECH, Korea, 2023

Academic Services

Conference reviewer: ECCV, ICCV [Outstanding Reviewer in 2025], NeurIPS, ICLR, CVPR, AAAI, BMVC, WACV
Journal reviewer: TVCJ

Template adapted from Jon Barron's website.