KupynOrest/epipolar-dpo

Official repo for: Epipolar Geometry Improves Video Generation Models

38
/ 100
Emerging

This project helps researchers and developers working with video generation models to create more realistic and geometrically consistent videos. It takes existing video outputs from a foundation model and applies epipolar geometry constraints using preference-based optimization to produce videos with improved 3D consistency and motion stability. This tool is for computer vision scientists, AI researchers, and engineers focused on advancing video synthesis technology.

Use this if you are generating videos with AI and need to significantly improve their 3D consistency and reduce geometric artifacts, especially in scenarios involving camera motion.

Not ideal if you are looking for a simple, out-of-the-box video generation tool for casual use without needing to delve into model fine-tuning or geometric consistency metrics.

video-synthesis computer-vision generative-AI 3D-consistency AI-model-refinement
No Package No Dependents
Maintenance 6 / 25
Adoption 9 / 25
Maturity 13 / 25
Community 10 / 25

How are scores calculated?

Stars

81

Forks

7

Language

Python

License

MIT

Last pushed

Oct 28, 2025

Commits (30d)

0

Get this data via API

curl "https://pt-edge.onrender.com/api/v1/quality/diffusion/KupynOrest/epipolar-dpo"

Open to everyone — 100 requests/day, no key needed. Get a free key for 1,000/day.