On Usability of Overcomplete Representations in Signal Analysis and Modeling

M. Namba and Y. Ishida (Japan)


Signal Analysis, PCA, Overcomplete Representation


A signal can be modeled by a linear combination of some unique waveforms. This paper discusses the usability of the overcomplete representations in the estimate of such waveforms, namely principal components. The experiment on quasi-periodic signal has shown that the principal components estimated by our specific procedure can better describe the signal, compared to the same number of components that are estimated by the conventional principal component analysis method.

Important Links:

Go Back