Emmanuel d'Angelo

http://www.computersdontsee.net

@sansuiso

[email protected]

The code presented here is a companion to my website. My goal is to provide some example code and tutorial-style explanations about variational image processing techniques.

While most of the related resources use matlab code, my goal is to provide C++ examples instead.

Modified BSD-style (see file LICENCE.txt)

• OpenCV 2.4, C++ API (but should work with previous 2.x C++ versions)
• cmake to build the library and the executables

The code repository is divided in two parts : a library and some command line applications.

• Rudin-Osher-Fatemi (TV-L2) denoising Implemented using algorithm 1 of [Ref. 1][1], i.e. primal-dual first order scheme without acceleration.

• TV constrained inpainting

[1]: Chambolle, A., Pock, T. (2010). A First-Order Primal-Dual Algorithm for Convex Problems with Applications to Imaging. Journal of Mathematical Imaging and Vision, 40(1), 120–145.