richzhang/PerceptualSimilarity

LPIPS metric. pip install lpips

47
/ 100
Emerging

This project helps image professionals accurately compare how similar two images or image patches are, moving beyond simple pixel-by-pixel comparisons. You provide it with two images, and it outputs a score reflecting their perceptual similarity as judged by humans. Anyone working with image quality assessment, image generation, or image processing who needs a reliable way to quantify visual differences would find this tool useful.

4,185 stars. No commits in the last 6 months.

Use this if you need to objectively quantify how visually similar two images or image modifications are, aligning with human perception.

Not ideal if you only need basic, mathematically exact pixel differences like Euclidean distance, or if your application does not involve visual perception.

image-quality-assessment computer-vision image-processing perceptual-evaluation generative-art
Stale 6m No Package No Dependents
Maintenance 0 / 25
Adoption 10 / 25
Maturity 16 / 25
Community 21 / 25

How are scores calculated?

Stars

4,185

Forks

523

Language

Python

License

BSD-2-Clause

Last pushed

Jul 02, 2024

Commits (30d)

0

Get this data via API

curl "https://pt-edge.onrender.com/api/v1/quality/ml-frameworks/richzhang/PerceptualSimilarity"

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