Depth profiling of thermally inhomogeneous materials by neural network recognition of photothermal time domain data

Authors
Glorieux, C Voti, RL Thoen, J Bertolotti, M Sibilia, C
Citation
C. Glorieux et al., Depth profiling of thermally inhomogeneous materials by neural network recognition of photothermal time domain data, J APPL PHYS, 85(10), 1999, pp. 7059-7063
Citations number
16
Categorie Soggetti
Apllied Physucs/Condensed Matter/Materiales Science
Journal title
JOURNAL OF APPLIED PHYSICS
ISSN journal
00218979 → ACNP
Volume
85
Issue
10
Year of publication
1999
Pages
7059 - 7063
Database
ISI
SICI code
0021-8979(19990515)85:10<7059:DPOTIM>2.0.ZU;2-N
Abstract
Thermal conductivity depth profiles of thermally inhomogeneous materials ar e retrieved from the time dependence of the surface temperature after a fla sh illumination. A neural network method, which is trained to recognize the correlation between depth profiles and the surface temperature on the basi s of many examples, is employed. Depth profiles retrieved from simulated no isy signals are shown and the average reconstruction errors are analyzed fo r different circumstances. (C) 1999 American Institute of Physics. [S0021-8 979(99)07710-5].