devzhk/InverseBench

InverseBench (ICLR 2025 spotlight)

49
/ 100
Emerging

InverseBench helps scientists working with challenging physical data, such as medical images or seismic readings, to accurately reconstruct the original information from incomplete or noisy observations. It takes the observed data and an algorithm choice, then outputs improved reconstructions and detailed performance metrics specific to the scientific domain. This tool is for researchers and practitioners in fields like medical imaging, astronomy, or geophysics who need to evaluate and compare advanced image reconstruction methods.

105 stars.

Use this if you are a scientist or engineer who needs to benchmark and understand the performance of diffusion-prior-based algorithms for inverse problems in physical sciences, like reconstructing black hole images or analyzing fluid dynamics from partial data.

Not ideal if you are looking for a ready-to-use application for general image restoration or if your inverse problem doesn't involve complex scientific data from fields like tomography or seismology.

medical imaging black hole imaging seismology fluid dynamics optical tomography
No Package No Dependents
Maintenance 10 / 25
Adoption 9 / 25
Maturity 16 / 25
Community 14 / 25

How are scores calculated?

Stars

105

Forks

12

Language

Python

License

MIT

Last pushed

Feb 10, 2026

Commits (30d)

0

Get this data via API

curl "https://pt-edge.onrender.com/api/v1/quality/diffusion/devzhk/InverseBench"

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