lixiaowen-xw/DiffuEraser
A diffusion-based video inpainting model that fills masked regions in videos while preserving temporal consistency.
Not currently ranked — collecting fresh signals.
star history
DiffuEraser is a diffusion model designed for video inpainting tasks, capable of filling masked or missing regions in video sequences. It leverages a diffusion architecture to generate coherent content that matches surrounding frames while maintaining temporal consistency across the video. The model aims to outperform prior approaches like Propainter in both content completeness and temporal consistency, while keeping computational efficiency acceptable. Inference and training code are publicly released.
Frequently asked
- What is lixiaowen-xw/DiffuEraser?
- A diffusion-based video inpainting model that fills masked regions in videos while preserving temporal consistency.
- Is DiffuEraser open source?
- Yes — lixiaowen-xw/DiffuEraser is open source, released under the Apache-2.0 license.
- What language is DiffuEraser written in?
- lixiaowen-xw/DiffuEraser is primarily written in Python.
- How popular is DiffuEraser?
- lixiaowen-xw/DiffuEraser has 676 stars on GitHub.
- Where can I find DiffuEraser?
- lixiaowen-xw/DiffuEraser is on GitHub at https://github.com/lixiaowen-xw/DiffuEraser.