Skip to content

OmniRe

Rebuild high-fidelity dynamic city scenes from device logs

Visit official site

What OmniRe is

The method is described as a way to reconstruct high-fidelity dynamic city scenes from device logs, using a Gaussian-based dynamic neural scene graph plus local canonical spaces for different actors.

What you can do with it

  • Reconstruct objects across a city scene
  • Model rigid motion of vehicles
  • Fit near pedestrians with a template SMPL model

Getting started

  • Visit the product page and read the overview
  • Download and install the required software
  • Prepare device logs of city scenes

Pros & cons

✓ What we like

  • Covers vehicles, pedestrians and cyclists
  • Supports simulation with all actors engaged

! What to watch out for

  • Input is described as device logs of city scenes
  • The evaluation cited is on the Waymo dataset

FAQ

Which actors are modelled?

The material names vehicles, pedestrians and cyclists.

What data is used?

The how-to section says to prepare device logs of city scenes.

Last reviewed: 2026-11-15

More AI image generation tools

View all →

How we review