Skip to content

DualGS

Dual Gaussian splatting for volumetric human video

Visit official site

What DualGS is

The method is described as a Gaussian-based volumetric video representation that optimizes joint and skin Gaussians for human performance.

What you can do with it

  • Render volumetric video in real time on modest devices
  • Capture complex human motion from multi-view input
  • Compress volumetric video with a hybrid strategy

Application scenarios

  • A concert performer rendered live in a VR headset
  • Athlete motion capture shown on a mobile device
  • A film effects preview of human movement

Pros & cons

✓ What we like

  • Multi-view input captures challenging human performance
  • Joint and skin Gaussians support robust tracking

! What to watch out for

  • Rendering settings must be tuned per device
  • Software or plugins are installed before use

FAQ

Where was it shown?

The material says SIGGRAPH Asia 2024.

What compression is stated?

Up to 120x with a hybrid compression strategy.

Last reviewed: 2026-11-15

More AI video generation tools

View all →

How we review