✨ VPI Lab has Launched!

Visual & Physical Intelligence Lab.

About Us

We Build Computer Vision that Moves the World.

Our lab bridges the gap between digital perception and real-world interaction. Driven by advanced Computer Vision and Machine Learning & Robot Vision, our research equips machines to truly understand complex real-world environments. Our research spans 3D Reconstruction for precise geometric scene understanding, Social Kinematics for modeling human-agent dynamics, and LLMs & Creative AI for generating physics-based content. Through these core technologies, we build intelligence that perceives, reasons, and interacts with the world.

Team collaboration
Digital workspace

News

Our Latest Updates
 Mar2026 The VPI Lab is officially open at DGIST EECS!
 Jun2025 One paper on visual prompting for zero-shot depth completion is accepted at ICCV 2025!
 Jun2025 One paper on vision-language model-based trajectory prediction is accepted at IEEE TPAMI (IF 20.8)!
Feb2025 One paper on crowd behavior generation is accepted at CVPR 2025!
Sep2024 One paper on 4D Gaussian splatting is accepted at NeurIPS 2024!
 Jul2024 One paper on text-guided video diffusion models is accepted at ECCV 2024!
Feb2024 Two papers on language and diffusion models are accepted at CVPR 2024!
 Jan2024 One paper on unbounded NeRF is accepted at ICLR 2024!
 Jul2023 One paper on low-rank approximated trajectory descriptors is accepted at ICCV 2023!
Apr2023 One paper on hyperbolic spatial propagation is accepted at ICML 2023!
Nov2022 One paper on control point prediction and refinement network is accepted at AAAI 2023!
 Jul2022 One paper on pedestrian group representation learning is accepted at ECCV 2022!
Mar2022 One paper on non-probabilistic sampling network is accepted at CVPR 2022!
Dec2020 One paper on disentangling and multi-relational GCN is accepted at AAAI 2021!
Want to join us?

We Are Recruiting Students & Interns

We welcome highly motivated M.S./Ph.D. students and undergraduate interns. If you are interested in our lab, please send your CV, transcript, and research interests to Prof. Bae.