Three papers accepted to CVPR 2026
Selected work spans dynamic indoor scenes, robust pose estimation, and in-the-wild 3D reconstruction.
Stanford Civil & Environmental Engineering
Gradient Spaces Lab
We study how real-world visual data can help design, construct, and understand adaptive spaces that move between the physical and digital world.
About
Learning spatial futures from the world as it is.
Gradient Spaces develops quantitative methods for environments that blend real reality, mixed reality, and virtual reality. The lab works across visual data, 3D understanding, design, construction, and human experience.
Our research spans dynamic 3D scene understanding, in-the-wild reconstruction, mixed reality, and sustainable built environments — with open datasets, code, and benchmarks released alongside our publications.
News
Latest signals
U. V. Helava Award for Best Paper
Tao Sun, Iro Armeni, and collaborators were recognized for work on evolving 3D scenes.
Two papers accepted to NeurIPS 2025
GuideFlow3D and Rectified Point Flow join the lab's growing archive of spatial AI work.
Research
Featured projects
CVPR 2026
ReScene4D
Temporally consistent semantic instance segmentation for evolving indoor 3D scenes.
CVPR 2026
WildPose
A unified framework for robust pose estimation in unconstrained visual settings.
CVPR 2026
GaussFusion
Geometry-informed video generation for stronger 3D reconstruction in the wild.
Connect