Title :
E.M. Inverse Scattering and Multi-Layer Perceptrons: Towards the automatic reconstruction of buried layers´ properties
Author :
Caorsi, Salvatore ; Stasolla, Mattia
Author_Institution :
Dept. of Electron., Univ. of Pavia, Pavia, Italy
Abstract :
The goal of this paper is a preliminary robustness assessment of a recently published ANN-based algorithm for the evaluation of subsurface layers´ properties. In particular, the analysis will focus on the dependency of overall performances on the training set dimensions and the neural networks capabilities of managing numerical errors.
Keywords :
buried layers; electromagnetic wave scattering; ground penetrating radar; multilayer perceptrons; neural nets; ANN-based algorithm; EM inverse scattering; automatic reconstruction; buried layers property; electromagnetic wave scattering; multilayer perceptrons; neural network capability; robustness assessment; subsurface layer property; training set dimension; Artificial neural networks; Data models; Error analysis; Ground penetrating radar; Permittivity; Robustness; Training;
Conference_Titel :
Electromagnetic Theory (EMTS), 2010 URSI International Symposium on
Conference_Location :
Berlin
Print_ISBN :
978-1-4244-5155-5
Electronic_ISBN :
978-1-4244-5154-8
DOI :
10.1109/URSI-EMTS.2010.5637142
