dle: Dictionary learning tools for Matlab. Dictionary Learning is a topic in the Signal Processing area, the dictionary is usually used for Sparse Representation or Approximation of signals. A dictionary is a collection of atoms, here the atoms are real column vectors of length N. A finite dictionary of K atoms can be represented as a matrix D of size NxK. In a Sparse Representation a vector x is represented or approximated as a linear combination of some few of the dictionary atoms. The approximation xa can be written as xa = D w where w is a vector containing the coefficients and most of the entries in w are zero. Dictionary Learning is the problem of finding a dictionary such that the approximations of many vectors, the training set, are as good as possible given a sparseness criterion on the coefficients, i.e. allowing only a small number of non-zero coefficients for each approximation. This page describes some experiments done on Dictionary Learning. The complete theory of dictionary learning is not told here, only a brief overview (of some parts) is given in section 3. and some links to relevant papers are included on the upper right part of this page. Section 4 presents the results of the experiments used in the RLS-DLA paper, and section 6 also includes the files needed to redo the experiments.
Keywords for this software
References in zbMATH (referenced in 1 article )
Showing result 1 of 1.
- Irannejad, Maziar; Mahdavi-Nasab, Homayoun: Low bit-rate SNR scalable video coding based on overcomplete dictionary learning and sparse representation (2020)