ASCL.net

Astrophysics Source Code Library

Making codes discoverable since 1999

ASCL Code Record

[ascl:2401.016] CRR: Convex Ridge Regularizer

CRR (Convex Ridge Regularizer) takes the gradient of regularizers that are the sum of convex-ridge functions and parameterizes them using a neural network that has a single hidden layer with increasing and learnable activation functions. The neural network is trained within a few minutes as a multistep Gaussian denoiser, and offers improvements for denoising and image reconstruction over other methods with similar reliability.

Code site:
https://github.com/axgoujon/convex_ridge_regularizers
Described in:
https://ui.adsabs.harvard.edu/abs/2022arXiv221112461G
Bibcode:
2024ascl.soft01016G

Views: 189

ascl:2401.016
Add this shield to your page
Copy the above HTML to add this shield to your code's website.