ECCV 2026
Extending a Large View Synthesis Model for Multi-view Panoptic Segmentation
Kwonyoung Ryu, In-Jae Lee, Jonghyun Jin, Hyunjee Lee, Jongmin Lee, Jaesik Park
Kwonyoung Ryu
I am a Ph.D. candidate in the Graduate School of Artificial Intelligence at POSTECH, advised by Prof. Sunghyun Cho and Prof. Jaesik Park. Currently, I am on a research assignment at the Visual & Geometric Intelligence Lab, Seoul National University, with Prof. Jaesik Park. Previously, I received my M.S. degree in Mechanical Engineering at Visual Intelligence Lab, KAIST, where I was advised by Prof. Kuk-Jin Yoon.
My research focuses on robust 3D scene understanding in challenging real-world settings, including 3D perception within multi-view image reconstruction and camera-LiDAR perception for autonomous driving.
News
Jun 2026
Our paper on extending a large view synthesis model for multi-view panoptic
segmentation was accepted to ECCV 2026.
Dec 2025
OpenBox was selected as a
NeurIPS 2025 Spotlight.
Dec 2025
Received the Top Reviewer Award at NeurIPS 2025.
Education
2021 - Present
Ph.D. in Graduate School of Artificial Intelligence | POSTECH
- 2023.09 - Present Visual & Geometric Intelligence Lab | SNU
- 2021.03 - 2023.08 Computer Vision Lab | POSTECH
2019 - 2021
M.S. in Mechanical Engineering | KAIST
- 2019.03 - 2021.02 Visual Intelligence Lab | KAIST
Thesis: Scanline Resolution-Invariant Depth Completion Using a Single Image and Sparse LiDAR Point Cloud
Advisor: Kuk-Jin Yoon
2014 - 2019
B.S. in Mechanical Engineering | KAIST
Publications
NeurIPS 2025
OpenBox: Annotate Any Bounding Boxes in 3D
In-Jae Lee, Mungyeom Kim, Kwonyoung Ryu, Pierre Musacchio, Jaesik Park
Spotlight
arXiv 2025
SpatialMosaic: A Multiview VLM Dataset for Partial Visibility
Kanghee Lee, In-Jae Lee, Minseok Kwak, Jungi Hong, Kwonyoung Ryu, Jaesik Park
CVPR 2023
Instant Domain Augmentation for LiDAR Semantic Segmentation
Kwonyoung Ryu*, Soonmin Hwang*, Jaesik Park
arXiv 2021
Revisiting LiDAR Registration and Reconstruction: A Range Image Perspective
Wei Dong*, Kwonyoung Ryu*, Michael Kaess, Jaesik Park
RA-L 2021
Scanline Resolution-invariant Depth Completion using a Single Image and Sparse LiDAR Point Cloud
Kwonyoung Ryu, Kang-il Lee, Jegyeong Cho, Kuk-Jin Yoon
RA-L 2020
Loop-Net: Joint Unsupervised Disparity and Optical Flow Estimation of Stereo Videos with Spatiotemporal Loop Consistency
Taewoo Kim, Kwonyoung Ryu, Kyeongseob Song, Kuk-Jin Yoon
Academic Service
- Conference Reviewer: CVPR (23-26), ECCV (22, 24, 26), NeurIPS (25-26), BMVC (26)
- Journal Reviewer: IEEE Robotics and Automation Letters (RA-L) (22-26)
- Top Reviewer Award, NeurIPS 2025