/haardict

Haar-like dictionary learning method

Primary LanguagePythonGNU General Public License v3.0GPL-3.0

Introduction

This repository contains python code used for testing the Haar dictionary learning method proposed in my PhD thesis.

Dependencies

The following python packages are used:

numpy
scipy
pywt
skimage

External Files and Optional Dependencies

To compute the KSVD a working installation of octave and oct2py are needed, as well as the KSVD-Box [1] package which must be compiled. To compute the HaarPSI values of the figures the haarpsi.py file is needed [2]. To use spectral clustering with the EMD (Earth Mover's Distance) the python pyemd package is needed.

Example Usage

For image reconstruction tasks the user will mostly be interested in the Test class. See file test.py for example usage.

[1] http://www.cs.technion.ac.il/~ronrubin/software.html

[2] http://www.haarpsi.org/