Time-Frequency Learning Machines

The theory of reproducing kernels has made a major breakthrough in the field of pattern recognition. It has led to new algorithms, with improved performance and lower computational cost, for non-linear analysis in high dimensional feature spaces.

These works extend the framework of the kernel learning machines to time-frequency analysis, showing that some specific reproducing kernels allow these algorithms to operate in the time-frequency domain. This link offers new perspectives in the field of non-stationary signal analysis, which can benefit from the developments of pattern recognition and Statistical Learning Theory.

Figure: The most relevant eigen-distributions in the time-frequency domain of K-complex signals in EEG recordings.

Selected Publications

P. HONEINE, C. RICHARD, and P. FLANDRIN
Nonstationary signal analysis with kernel machines
chapter in Handbook of Research on Machine Learning Applications and Trends: Algorithms, Methods and Techniques, Publisher: Information Science Reference, IGI Global, ISBN: 1605667668, pp 223-241 2010.

P. HONEINE, C. RICHARD, and P. FLANDRIN
Time-frequency learning machines
IEEE Trans. Signal Processing, 55:3930-3936, 2007.

P. HONEINE, and C. RICHARD
Distribution temps-fréquence à paramétrisation radialement Gaussienne optimisée pour la classification
English: Optimizing time-frequency representations for signal classification using radially Gaussian kernels
Traitement du signal, 25(6), 2008. (invited french paper)

P. HONEINE, and C. RICHARD
Signal-dependent time-frequency representations for classification using a radially gaussian kernel and the alignment criterion
Proc. of IEEE Statistical Signal Processing Workshop (SSP), Madison (WI), USA, 26-29 Aug. 2007.
[paper] [poster] [bibTeX]

P. HONEINE, C. RICHARD, P. FLANDRIN, and J.-B. POTHIN
Optimal selection of time-frequency representations for signal classification: a kernel-target alignment approach
Proc. of IEEE International Conference on Acoustics, Speech and Signal Processing (ICASSP), Toulouse, France, 14-19 May 2006.
[paper] [poster] [bibTeX]

P. HONEINE, et C. RICHARD
Distribution temps-fréquence à noyau radialement gaussien : optimisation pour la classification par le critère d’alignement noyau-cible
Actes du XXI-ème colloque GRETSI : Traitement du signal et des images, Troyes, France, 2007.
[paper] [bibTeX]

P. HONEINE, C. RICHARD, et P. FLANDRIN
Reconnaissance des formes par méthodes à noyau dans le domaine temps-fréquence
Actes du XX-ème colloque GRETSI : Traitement du signal et des images, Louvain-la-neuve, Belgium, 2005.
[paper] [bibTeX]