Skip to content

Diffusion Priors

Synthesise new views from a monocular video

Visit official site

What Diffusion Priors is

The paper describes a diffusion prior approach to dynamic view synthesis that generates new views of a scene from monocular video.

What you can do with it

  • Fine-tune frames from a single video
  • Apply knowledge distillation in the process
  • Evaluate results on complex scenes

Method notes

  • Geometric consistency is listed as a goal
  • Scene consistency is listed alongside it
  • Qualitative and quantitative experiments are described

Pros & cons

✓ What we like

  • Both geometric and scene consistency are targeted
  • The method is assessed with qualitative and quantitative experiments

! What to watch out for

  • The input is described as monocular video
  • The work is presented as a paper rather than a product

FAQ

What does the method produce?

New views of dynamic scenes from monocular video.

What consistency is claimed?

The material names geometric consistency and scene consistency.

Last reviewed: 2026-11-15

More AI video generation tools

View all →

How we review