Skip to content

Deblurring 3D Gaussian Splatting

Rebuild sharp images from blurry input with real-time rendering

Visit official site

What Deblurring 3D Gaussian Splatting is

The framework is described as a neural field deblurring method that uses 3D Gaussians, rasterization and a small MLP.

What you can do with it

  • Rebuild clear detail from blurry images
  • Render a scene in real time
  • Add points to densify a sparse point cloud

Example uses

  • Handle medical images in a web app
  • Process images for film effects
  • Work as a virtual reality image plugin

Pros & cons

✓ What we like

  • A small MLP is used in the pipeline
  • 3D Gaussians are pruned by relative depth

! What to watch out for

  • The described output is a research framework
  • Training uses KNN to add points

FAQ

What does it reconstruct?

The material says detail-clear images are rebuilt from blurry images.

Is rendering real time?

Yes, real-time rendering is listed in the material.

Last reviewed: 2026-11-15

More AI image generation tools

View all →

How we review