aangelopoulos/im2im-uq
Image-to-image regression with uncertainty quantification in PyTorch. Take any dataset and train a model to regress images to images with rigorous, distribution-free uncertainty quantification.
This project helps researchers and engineers who work with medical images like MRI scans or electron microscopy. It allows you to train a system that takes a raw, noisy, or incomplete image and transforms it into a clear, reconstructed image, along with a crucial 'uncertainty map'. This map highlights areas where the reconstruction might be unreliable, giving you confidence in the results.
No commits in the last 6 months.
Use this if you need to transform one type of image into another (e.g., low-resolution to high-resolution, or noisy to clear) and absolutely require a quantifiable measure of how trustworthy each part of the output image is.
Not ideal if your task does not involve image-to-image conversion, or if you don't need a rigorous uncertainty measure for your image predictions.
Stars
59
Forks
10
Language
Python
License
MIT
Last pushed
Feb 23, 2023
Commits (30d)
0
Get this data via API
curl "https://pt-edge.onrender.com/api/v1/quality/diffusion/aangelopoulos/im2im-uq"
Open to everyone — 100 requests/day, no key needed. Get a free key for 1,000/day.