Action-Image MoT World Model — Rollouts, Short vs Long, Train vs Test
Explore video & tactile rollouts from an action‑conditioned model
Explore video & tactile rollouts from an action‑conditioned model
Generate multi‑view action images from robot pose data
Explore AI‑edited videos with object add/remove
Explore a three‑view 3D point‑cloud reconstruction online
Compare point‑cloud sampling methods interactively
Explore AI-edited videos with object addition and removal
Explore visuo‑tactile model rollouts and performance metrics
Downsample point clouds while preserving geometry
Explore AI-edited videos with objects added or removed
Compare audio separation quality across pipelines
Explore tactile model performance across conditions
Generate future tactile video frames from action inputs
Remove sounding objects from videos and enhance AV pairs
Remove a specific object and its sound from a video clip
Add or remove objects in videos with a text prompt
Explore model predictions of vision and touch
Generate per-pixel tactile motion maps from sensor actions
Explore tactile video predictions vs ground truth
Refine video quality with adjustable denoise strength
Compare original and reconstructed tactile videos side‑by‑side
Compare video reconstruction quality of two model adapters