Skip to content

CHOIS

Synthesize human and object motion together

Visit official site

What CHOIS is

CHOIS is described as a method that generates object motion and human motion together from language, initial states and sparse object waypoints.

What you can do with it

  • Describe an interaction in words
  • Supply waypoints for the object path
  • Generate matching human and object motion

How synthesis works

  • A conditional diffusion model drives the synthesis
  • Geometry loss aligns motion with the waypoints
  • Guidance terms enforce contact constraints

Pros & cons

✓ What we like

  • Language description drives both motions
  • Contact constraints are applied during sampling

! What to watch out for

  • Initial object and human states must be prepared
  • Waypoints have to be set before synthesis

FAQ

What inputs are needed?

A language description, initial states and sparse object waypoints.

What model is used?

A conditional diffusion model is described.

Last reviewed: 2026-11-15

More AI video generation tools

View all →

How we review