Skip to content

HyFluid

Infer fluid density and velocity from video

Visit official site

What HyFluid is

The method is described as a neural approach that recovers fluid density and velocity from sparse multi-view video.

What you can do with it

  • Recover a 3D fluid field from views
  • Visualize the restored density and speed
  • Re-simulate the flow from a new angle

Application scenarios

  • Reconstruct a field on a website project
  • Forecast later states of a flow
  • Synthesize a moving neural scene

Pros & cons

✓ What we like

  • Re-simulation from novel views is listed
  • Future prediction is a named use

! What to watch out for

  • Input is sparse multi-view video
  • The method targets incompressible flow

FAQ

What is inferred?

The material names 3D fluid density and velocity fields.

Which applications are listed?

Re-simulation, editing, future prediction and scene synthesis are named.

Last reviewed: 2026-11-15

More LLM API platform tools

View all →

How we review