From Neuronal cost-based metrics towards sparse coded signals classification

Abstract : Sparse signal decomposition are keys to efficient compression, storage and denoising, but they lack appropriate methods to exploit this sparcity for a classification purpose. Sparse coding methods based on dictionary learning may result in spikegrams, a sparse and temporal representation of signals by a raster of kernel occurrence through time. This paper proposes a method for coupling spike train cost based metrics (from neuroscience) with a spikegram sparse decompositions for clustering multivariate signals. Experiments on character trajectories, recorded by sensors from natural handwriting, prove the validity of the approach, compared with currently available classification performance in literature.
Complete list of metadatas

Cited literature [12 references]  Display  Hide  Download

https://hal.archives-ouvertes.fr/hal-00731497
Contributor : Jerome Mars <>
Submitted on : Thursday, February 27, 2014 - 3:44:41 PM
Last modification on : Monday, May 13, 2019 - 11:15:31 AM
Long-term archiving on : Tuesday, May 27, 2014 - 10:35:34 AM

File

2012_Mouraud_ESANN-1.pdf
Files produced by the author(s)

Identifiers

  • HAL Id : hal-00731497, version 1

Citation

Anthony Mouraud, Quentin Barthélemy, Aurélien Mayoue, Anthony Larue, Cedric Gouy-Pailler, et al.. From Neuronal cost-based metrics towards sparse coded signals classification. 20th European Symposium on Artificial Neural Networks, Computational Intelligence and Machine Learning (ESANN 2012), Apr 2012, Bruges, Belgium. pp.311-316. ⟨hal-00731497⟩

Share

Metrics

Record views

634

Files downloads

161