Chung-Ang University symbol Chung-Ang University

Spatial
Intelligence Lab

Dept. of Computer Science & Engineering, Chung-Ang University

We study computer vision, machine learning, robotics, and multi-modal LLMs, teaching machines to perceive, reason about, and act in the 3D world.

Research Fields

Our work spans six pillars that connect visual geometry, foundation models, and embodied AI.

01
Computer Vision & Visual Geometry
Recovering 3D structure from 2D images via correspondence, pose, geometry, reconstruction, and equivariance.
Visual CorrespondenceCamera & Object PoseMulti-View Geometry3D Reconstruction
02
Multi-Modal Large Language Models
Building multi-modal LLMs that see, hear, reason, and act across images, audio, video, and 3D scenes.
Multi-Modal Foundation ModelsChain-of-Thought ReasoningPost-Training (DPO, GRPO)3D-Aware MLLMs
03
Geometric Representation Learning
Symmetry, invariance, and equivariance as inductive biases for geometrically aware representations.
Invariance & EquivarianceSO(3) / Spherical HarmonicsSelf-Supervised LearningGeometric Deep Learning
04
3D Visual Foundation Models
Feed-forward 3D foundation models that recover geometry, correspondence, and renderable scenes from images.
DUSt3R / VGGTPoint Maps & 3D GaussiansFeed-Forward ReconstructionNovel-View Synthesis
05
Robotics for Real-World Physical AI
Building embodied agents that perceive in 3D, reason with language, and act reliably in the physical world.
Embodied AIVision-Language-ActionSim-to-RealManipulation
06
Application Research, AI+X
Spatial AI as a core engine for AI+X across AR/VR, driving, healthcare, infrastructure, and animal behavior.
AR / VRAutonomous DrivingMedical ImagingAnimal Behavior

Recent updates from the Spatial Intelligence Lab.