Skip to content

Make-Your-Anchor

A diffusion framework that turns one minute of video into anchor-style clips

Visit official site

What Make-Your-Anchor is

The framework is described as a diffusion-model-based 2D virtual avatar generation framework that builds anchor-style videos from about one minute of source footage. It reproduces precise upper-body and hand motions and adds an identity-specific face enhancement module.

What you can do with it

  • Generate anchor-style videos from about one minute of video material
  • Reproduce upper-body and hand motions with precision
  • Render 3D mesh states into human appearance with a structure-guided diffusion model
  • Generate videos of arbitrary length through batch-overlapped temporal denoising

Model notes

  • A two-stage training strategy binds motion to a specific appearance
  • The frame-wise 2D U-Net is extended to a 3D form
  • Batch-overlapped temporal denoising lifts the video length limit at inference
  • An identity-specific face enhancement module improves the face area

Pros & cons

✓ What we like

  • Structure-guided diffusion renders 3D mesh states into human appearance
  • A two-stage training strategy binds motion to a specific appearance

! What to watch out for

  • The described input is a video clip of about one minute
  • Rendering runs through 3D mesh states and a two-stage training strategy

FAQ

How much footage is needed?

The material states about one minute of video material is enough.

What do the reported experiments cover?

The description states experiments on visual quality, temporal consistency and identity fidelity that reported advantages over existing techniques.

Last reviewed: 2026-11-15

More AI video generation tools

View all →

How we review