PaulCouairon/DiffCut

[NeurIPS 2024] Official code for DiffCut: Catalyzing Zero-Shot Semantic Segmentation with Diffusion Features and Recursive Normalized Cut

36
/ 100
Emerging

DiffCut helps automatically outline distinct objects or regions within images, even for categories it hasn't been explicitly trained on. You provide an image, and it outputs a segmented version with different areas highlighted. This tool is useful for researchers and practitioners working in computer vision, autonomous systems, or medical imaging who need to precisely delineate objects in a wide variety of images without extensive manual labeling.

No commits in the last 6 months.

Use this if you need to precisely segment images into meaningful regions for objects or concepts you haven't pre-defined, saving significant manual annotation time.

Not ideal if you require segmentation of extremely fine-grained details or highly ambiguous objects that even humans struggle to differentiate.

computer-vision image-analysis zero-shot-learning object-segmentation autonomous-vision
Stale 6m No Package No Dependents
Maintenance 0 / 25
Adoption 8 / 25
Maturity 16 / 25
Community 12 / 25

How are scores calculated?

Stars

50

Forks

6

Language

Jupyter Notebook

License

MIT

Last pushed

Jan 19, 2025

Commits (30d)

0

Get this data via API

curl "https://pt-edge.onrender.com/api/v1/quality/diffusion/PaulCouairon/DiffCut"

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