Hakjin Lee

Hakjin Lee

Research Engineer at PIT IN Corp.

I study how learning systems can reliably act in the physical world under limited experience. At PIT IN Corp., my research centers on 3D vision for robotics — primarily stereo vision and object pose estimation — and on robot learning, motivated by the goal of building real products and services that interact with the physical world.

Previously, I was at SI Analytics and Satrec Initiative, where I worked on remote sensing computer vision: synthetic data augmentation for change detection, rotated detection transformers, and spatiotemporal models for weather forecasting. I also contribute to the OpenMMLab open-source ecosystem.

hj at pitin-ev dot com  •  Google Scholar  •  GitHub

Updates

Research

Distributional Stability
Distributional Stability of Tangent-Linearized Gaussian Inference on Smooth Manifolds

Junghoon Seo, Hakjin Lee, Jaehoon Sim
IEEE Robotics and Automation Letters (RA-L), 2026
Paper  •  Code

YOPO
You Only Pose Once: A Minimalist's Detection Transformer for Monocular RGB Category-level 9D Multi-Object Pose Estimation

Hakjin Lee, Junghoon Seo, Jaehoon Sim
IEEE International Conference on Robotics and Automation (ICRA), 2026
Project  •  Paper  •  Code  •  Video

RHINO
Hausdorff Distance Matching with Adaptive Query Denoising for Rotated Detection Transformer (RHINO)

Hakjin Lee*, MinKi Song*, Jamyoung Koo, Junghoon Seo  (* equal contribution)
IEEE/CVF Winter Conference on Applications of Computer Vision (WACV), 2025
Paper  •  Code

Masked Autoregressive Weather
Masked Autoregressive Model for Weather Forecasting

Doyi Kim, Minseok Seo, Hakjin Lee, Junghoon Seo
arXiv preprint arXiv:2409.20117, 2024
Paper  •  Code

Self-pair
Self-pair: Synthesizing Changes from Single Source for Object Change Detection in Remote Sensing Imagery

Minseok Seo, Hakjin Lee, Yongjin Jeon, Junghoon Seo
IEEE/CVF Winter Conference on Applications of Computer Vision (WACV), 2023
Paper  •  Code

Selected Projects

Experience

Awards

Education