RESEARCH
Perception that holds up outside the lab.
The DROP Lab started underwater, building 3D maps of reefs, shipwrecks, and submerged cities from autonomous vehicles. The same questions — how a robot builds a faithful model of the world from imperfect sensors, and how it acts safely on that model — now drive our work in manipulation and on land.
01
Reliable manipulation
Failure detection, recovery, and motion generation for learned visuomotor policies.
02
3D scene understanding
Reconstruction, segmentation, and depth from images — and knowing when geometry is ambiguous.
03
Underwater robotics
AUV mapping, seafloor reconstruction, color correction, and generative underwater terrain.
- Infinite Leagues Under the Sea: Photorealistic 3D Underwater Terrain Generation by Latent Fractal Diffusion Models2025
- RecGS: Removing Water Caustic with Recurrent Gaussian Splatting2024
- Beyond NeRF Underwater: Learning Neural Reflectance Fields for True Color Correction of Marine Imagery2023
- All in this theme →
04
Autonomy in the field
Perception for self-driving and delivery robots operating among people.
- Unified Self-supervised Representation Learning for Multi-modal and Single-modal 3D Perception2025
- DOSE3 : Diffusion-based Out-of-distribution detection on SE(3) trajectories2025
- Scalable Benchmarking and Robust Learning for Noise-Free Ego-Motion and 3D Reconstruction from Noisy Video2025
- All in this theme →