Stars
[ECCV 2026] Feed-forward, Physically Plausible Human-Scene Reconstruction
Official code and data from DexWM ("World Models Can Leverage Human Videos for Dexterous Manipulation").
RAIDEN is a toolkit for YAM robots that streamlines calibration, coordinated bimanual data collection, and dataset conversion.
[ECCV 2026] Official implementation of paper LiveWorld: Simulating Out-of-Sight Dynamics in Generative Video World Models
Official repository for the paper "Vision Transformers for End-to-End Vision-Based Quadrotor Obstacle Avoidance" by Bhattacharya, et al. (2024) from GRASP, Penn.
EgoDex: Learning Dexterous Manipulation from Large-Scale Egocentric Video
[SIGGGRAPHASIA2025] PhySIC: Physically Plausible 3D Human-Scene Interaction and Contact from a Single Image
GentleHumanoid: Whole body Motion Tracking with Compliance - Inference and Deploy
code release of paper "DIMOS: Synthesizing Diverse Human Motions in 3D Indoor Scenes"
official implementation of [PRIMAL: Physically Reactive and Interactive Motor Model for Avatar Learning, ICCV'25]
[CVPR 2025] Vid2Sim: Generalizable, Video-based Reconstruction of Appearance, Geometry and Physics for Mesh-free Simulation
[ICLR 2026] An unified model for 4D human-scene reconstruction
HaWoR: World-Space Hand Motion Reconstruction from Egocentric Videos
Code for the paper MultiPhys: Multi-Person Physics-aware 3D Motion Estimation (CVPR 2024)
The Fast Way From Vertices to Parametric 3D Humans
[CVPR 2025] PICO: Reconstructing 3D People In Contact with Objects
Agent-to-Sim Learning Interactive Behavior from Casual Videos.
[CVPR25 Oral (Top 3.3%)] Official code for paper "Reconstructing Humans with a Biomechanically Accurate Skeleton".
Code for "GVHMR: World-Grounded Human Motion Recovery via Gravity-View Coordinates", Siggraph Asia 2024, TPAMI 2026
Code for our CVPR'25 paper - "DiffLocks: Generating 3D Hair from a Single Image using Diffusion Models"
Visual Imitation Enables Contextual Humanoid Control. CoRL 2025, Best Student Paper Award.