Hierarchical Radial Basis Functions Networks (HRBF) have been recently introduced as a tool for adaptive multiscale image reconstruction from range data. They are based on local operation on the data and are able to give a sparse approximation. In this paper HRBF are reframed for the regular sampling case, and they are compared with Wavelet Decomposition. Results show that HRBF, thanks to their constructive approach to approximation, are much more tolerant on errors on the parameters when errors occurs in the configuration phase, while they are more sensitive to the errors which occurs since the network has been configured.
Multi-resolution models for data processing : an experimental sensitivity analysis / S. Ferrari, N. A. Borghese, V. Piuri - In: Smart connectivity : integrating measurement and control : IMTC/2000 : proceedings of the 17th IEEE instrumentation and measurement technology conference : Baltimore, Maryland, USA, May 1-4, 2000 / [a cura di] [s. n.]. - Piscataway : Institute of Electrical and Electronics Engineers, 2000. - ISBN 0780358902. - pp. 1056-1060 (( Intervento presentato al 17. convegno IEEE Instrumentation and Measurement Technology Conference (IMTC) tenutosi a Baltimore nel 2000.
Multi-resolution models for data processing : an experimental sensitivity analysis
S. FerrariPrimo
;N. A. BorgheseSecondo
;V. PiuriUltimo
2000
Abstract
Hierarchical Radial Basis Functions Networks (HRBF) have been recently introduced as a tool for adaptive multiscale image reconstruction from range data. They are based on local operation on the data and are able to give a sparse approximation. In this paper HRBF are reframed for the regular sampling case, and they are compared with Wavelet Decomposition. Results show that HRBF, thanks to their constructive approach to approximation, are much more tolerant on errors on the parameters when errors occurs in the configuration phase, while they are more sensitive to the errors which occurs since the network has been configured.
Pubblicazioni consigliate
I documenti in IRIS sono protetti da copyright e tutti i diritti sono riservati, salvo diversa indicazione.
