Highly Nonlinear Approximations for Sparse Signal Representation
Construction of Wavelet Dictionaries for ECG Signal Modeling
Algorithms and MATLAB implementation facilitating the construction of wavelet dictionaries for ECG modeling are presented. The model enables the representation of an ECG record as a linear combination of fewer elementary components than those required by a wavelet basis. That result was demonstrated in a previous work on the MIT-BIH Arrhythmia data set. This paper complements the previous one. It produces the technical details of the methods and algorithms which are implemented by the software for constructing different families of wavelet dictionaries. The implementation allows for straightforward further extensions to include additional wavelet families.
by Dana Cerna and Laura Rebollo-Neira
This project is supported by EPSRC (EP/D062632/1)
