Structured Intelligence Lab
We study reliable and efficient learning methods for structured visual data, with a focus on 3D vision, medical imaging, and trustworthy machine learning.
College of AI Convergence, Korea Aerospace University
Research Topics
3D Representation and Reconstruction
Neural representation, surface reconstruction, 3D generation, and neural rendering.
Medical Image Analysis
Deformable registration, multimodal alignment, and missing-data imputation.
Trustworthy Machine Learning
Robustness, uncertainty, bias mitigation, and learning under distribution shift.
News
Sungjin joined STIL Lab as an undergraduate research intern.
Our work on robust neural surface reconstruction was accepted to BMVC 2026.
Seungbin and Sunho joined STIL Lab as undergraduate research interns.
Our work on bias alleviation through network pruning was accepted to IEEE Transactions on Image Processing.
Our geometry-preserving robust neural reconstruction paper was accepted to the CVPR 2026 SPAR-3D Workshop.
Minjae, Seonwoo, Seoyoung, and Jihun joined STIL Lab.
Prof. Shin visited MBZUAI and gave a talk titled “Learning What Matters: Reliability-Aware 3D Vision.”
Jinju, Chaeyun, and Seungjun joined STIL Lab.
Our work on group-wise flat minima was accepted to Mathematics.
STIL Lab was established at the College of AI Convergence, Korea Aerospace University.
Lab Gallery
Photos can be updated later
Add lab activity photographs here when the final images are ready. The page structure will not change when photographs are replaced.