Graph Dictionary Learning

Yael Yankelevvsky’s work on handling graph-based signals has been reported in THIS paper. The ideas is to take into account the Laplacian matrix of the graph signals, and extend the dictionary leanring to accomodate for it. Indeed, our work has two additional ideas incorporated within this scheme: (1) we are able to learn the Laplacian matrix within the diciotnary learning process; and (ii) we take into account another Laplacian – ones that accounts for interrelations between the given different signals, thus turning our pursuit int a joint one. An accompanying package is provided to reproduce all the results shown in this paper.