A highly optimized, parallel implementation of the Batch-OMP version of the KSVD learning algorithm.
Switch branches/tags
Nothing to show
Clone or download
hoytak Merge pull request #2 from glemaitre/master
Just update the __init__.py to be able to use KSVD_Encode
Latest commit 0a3ff2c Nov 23, 2015
Permalink
Type Name Latest commit message Commit time
Failed to load latest commit information.
ksvd Fix the include for eigen Aug 21, 2015
tests Initial Commit. Jul 17, 2013
third-party Include eigen dependencies Aug 21, 2015
.gitignore Initial commit Jul 17, 2013
.gitmodules Include eigen dependencies Aug 21, 2015
LICENSE Initial commit Jul 17, 2013
README.md Update the README Aug 23, 2015
setup.py Initial Commit. Jul 17, 2013

README.md

pyksvd

Installation

Dependencies

  • Eigen (C++)
  • gcc >= 4.6 (C++)
  • numpy
  • scipy
  • cython

Installation

Eigen library is part of the repository as submodule in third-party. You need to:

  1. Init the submodule git submodule init at the root of the repository.
  2. Update the submodule git submodule update at the root of the repository.

Then, the installation can be performed by running python setup.py install.

About

A highly optimized, parallel implementation of the Batch-OMP version of the KSVD learning algorithm. It implements the algorithm in the paper. The computation is done in highly optimized C++ code with OpenMP implementations for multicore archetectures.

"Efficient Implementation of the K-SVD Algorithm and the Batch-OMP Method", by Ron Rubinstein, Michael Zibulevsky and Michael Elad, 2009.