File Exchange

image thumbnail


version (47.9 KB) by Torbjorn Ringholm
Image denoising/inpainting with Euler's elastica prior.


Updated 06 Oct 2017

GitHub view license on GitHub

DG-EulerElastica is a package for variational image analysis using Euler's elastica as a regularizer. It is designed to be easy to use, with little user input required, yet still providing good results. The underlying optimization algorithm used is based on discrete gradients from geometric numerical integration, and converges in all but the most extreme cases. Get started by downloading the package and running/reading the elasticaDemo.m script.
Three implementations are included. One (slow) in pure MATLAB code, one (faster) using the MEX interface with a C code for critical parts, and one (fastest) using MEX and parallelization.

The MATLAB Image Processing Toolbox is needed for the demo script only, and the Parallel Computing Toolbox for the parallel implementation only.

Cite As

Torbjorn Ringholm (2020). DG-EulerElastica (, GitHub. Retrieved .

Comments and Ratings (0)


Corrected typos

Updated algorithm image and description.

MATLAB Release Compatibility
Created with R2017a
Compatible with any release
Platform Compatibility
Windows macOS Linux