K-SVD - An Algorithm for Designing Overcomplete Dictionaries for Sparse Representation

PyKSVD is a Python implementation of the K-SVD algorithm based on paper K-SVD by Michal Aharon, Michael Elad, and Alfred Bruckstein.

  • Course: Machine Learning for Time Series led by Prof. Laurent Oudre

  • Code : GitHub
  • Report : PDF
  • Library : pip install pyksvd on PyPi

Project Overview

PyKSVD is a Python implementation of the K-SVD algorithm based on paper K-SVD: An Algorithm for Designing Overcomplete Dictionaries for Sparse Representation (2006) by Michal Aharon, Michael Elad, and Alfred Bruckstein.

Examples

  • Image processing : Denoise or complete a corrupted image with missing pixels with patches.
Image Reconstruction with 70% missing pixels