Auto-adjustable method for Gaussian width optimization on RBF neural network. Application to face authentication on a mono-chip system

Lionel Pierrefeu 1 Jacques Jay 1 Cécile Barat 1, *
* Corresponding author
1 LHC - Laboratoire Hubert Curien [Saint Etienne]
Abstract : This paper describes an automatic method for optimizing a radial basis function (RBF) neural network parameter during training stage. The neural network is used as a classifier to realize a human face authentication system. The aim of this project is to obtain a low cost system on chip (SoC) to replace password identification on mobile devices. The system is designed in order to respect the AAA methodology for algorithm selection and implantation on hardware platform. Several classifier parameters need to be adjusted to obtain better performances. One of these critical parameters is the Gaussian width. Contrary to other applications, there is a unique class (corresponding to the trained person); standard methods for width optimization therefore do not work. We suggest a new method to compute an optimized width.
Document type :
Conference papers
Complete list of metadatas
https://hal-ujm.archives-ouvertes.fr/ujm-00378472
Contributor : Cecile Barat <>
Submitted on : Friday, April 24, 2009 - 12:29:44 PM
Last modification on : Wednesday, November 20, 2019 - 3:13:27 AM

Identifiers

Collections

UNIV-ST-ETIENNE | CNRS | LAHC | IOGS | PARISTECH | UDL

Citation

Lionel Pierrefeu, Jacques Jay, Cécile Barat. Auto-adjustable method for Gaussian width optimization on RBF neural network. Application to face authentication on a mono-chip system. The 32nd Annual Conference of the IEEE Industrial Electronics Society Conservatoire National des Arts & Metiers, Nov 2006, Paris, France. pp.3481-3485, ⟨10.1109/IECON.2006.347848⟩. ⟨ujm-00378472⟩

Export

BibTeX TEI DC DCterms EndNote

Share

Metrics

Record views

79

Contact

Informations

CCSD