Dictionary Learning for Sparsely Representable Signals

Tutor: Tobias Schnier
Type of Thesis: Project (MSc)
date of end: 04/2016
Student: Tobias Monsees
Status: finished
ANT-shelfmark:
Abstract:

In this topic, the  Dictionary Learning problem is investigated.

Main questions:

What is the basis that lead to the signals present?

How can this basis mathematically be found using some given signals?

Application:

Imagine you here two people speak. How do you know there are two distinct signals? How can you distinguish between the two persons?

Work to do:

Literature search of already known algorithms for dictionary learning

MATLAB-programming of synthetically designed problems to evaluate the efficiency of the algorithms

Investigation of the problem of noise at the signal allocation

Expansion of given algorithms to be robust to noise

Expertise needed:

Basic mathematical background is needed

Last change on 31.05.2016 by
AIT ieee GOC tzi ith Fachbereich 1
© Department of Communications Engineering - University of BremenImprint / Contact