mohammadasim98/met3r

MEt3R: Measuring Multi-View Consistency in Generated Images

46
/ 100
Emerging

Computes differentiable multi-view consistency scores by leveraging DUSt3R for dense 3D reconstruction and warping image features across viewpoints, then comparing them using DINO embeddings upsampled via FeatUp. Supports multiple geometric backbones (MASt3R, DUSt3R, RAFT) and feature extractors (DINO, DiNOv2, CLIP, ResNet50) with configurable distance metrics (cosine similarity, LPIPS, PSNR, SSIM), enabling evaluation of novel view and video generation quality independent of the generative model used.

163 stars.

No Package No Dependents
Maintenance 10 / 25
Adoption 10 / 25
Maturity 16 / 25
Community 10 / 25

How are scores calculated?

Stars

163

Forks

10

Language

Python

License

MIT

Last pushed

Feb 23, 2026

Commits (30d)

0

Get this data via API

curl "https://pt-edge.onrender.com/api/v1/quality/diffusion/mohammadasim98/met3r"

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