Research method generating text-conditioned 3D human-object interaction motion using body-part anchors and a contact refiner.
Updated EU status checked Last shown
Sources: the vendor's own pages (linked above) and creator videos at the linked timestamps. EU status is information, not legal advice; see how we check.
PAMI (Part Anchored Motion for Text to Human-Object Interaction Generation) is a research framework that generates full-body human motion together with object trajectories from a text prompt and an object mesh. It was authored by Chuqiao Li, Xianghui Xie, Yong Cao, Andreas Geiger and Gerard Pons-Moll (2026).
The key idea, inspired by the Hough Transform, is to express object motion relative to several body-part anchors (root, wrists, ankles, plus a free anchor). PamiVAE learns an interaction latent space and decodes per-frame weights that aggregate the part-specific votes into the object trajectory. PamiGen produces a coarse interaction from text (CLIP conditioning) in this latent space, and PamiRefiner then iteratively resolves contact geometry using long-range body-part probes and short-range surface sensors near the hands.
On the InterAct benchmark, the authors report 14.5% higher contact recall than the previous state of the art, and compare qualitatively against LIGHT and InterAct, targeting object drift, missed contact and penetration. A YouTube creator reported the GitHub code is released under MIT with models to follow; no repository or weights URL was in the provided material. It is a research release, not a hosted product.
Watch the moment: the exact second top creators show it. Videos load only when you press play, via youtube-nocookie.com.
Generates 3D human animation of a person interacting with an object from a text prompt. Body parts vote on object motion and a refiner cleans up contact points. GitHub is out under MIT, with models to come.
Closest neighbours by meaning, not just by category.
Model
Unified motion-language model for text-to-motion, motion captioning, completion and editing, from HKUST and Tencent Games.
Research
Diffusion framework that learns text-to-3D human motion generation from 2D poses extracted from monocular videos.
Research
Research framework for first-person world simulation driven by 3D human motion and anchor-view scene customization.
Model · ByteDance
ByteDance unified video generation and editing framework pairing an MLLM semantic planner with a DiT renderer.
Model
Text-to-human-motion flow-matching model that generates full-resolution motion directly, with training-free joint constraints.
Research · ByteDance
ByteDance research pipeline that turns indoor captures into editable, simulation-ready 3D scenes of separate object meshes.