Improving the Counterpropagation network performances
- 24 Downloads
This paper deals with the problem of how input data normalization can affect the performances of the Counterpropagtion neural network. In the following, an example drawn from the landcover classification of remotely sensed images is presented and a solution, based on the Decorrelation Stretching technique, is proposed.
KeywordsNeural Network Artificial Intelligence Input Data Complex System Nonlinear Dynamics
Unable to display preview. Download preview PDF.
- P.D. Wasserman.Neural Computing, Theory and Practice, Van Nostrand Reinhold, New York, 1989.Google Scholar
- C. Bechini, L. Chiarantini, P. Ciotti, S. Moretti, E Pettinelli, N. Pierdicca. MAC'19 on Montespertoli: preliminary assessment of land polarimetric features,Proc. IGARSS, vol. 1, pp. 395–397, 1992.Google Scholar
- P. Coppo, P. Ferrazzoli, G. Luzi, S. Paloscia, G. Schiavon, C. Susini. MAC'91 on Montespertoli: preliminary analysis of multifrequency SAR sensitivity to soli and vegetation paramenters,Proc. IGARSS, Vol. 1, pp. 489–491, 1992.Google Scholar
- G. G. Wilkinson. The processing and interpretation of remotely-sensed satellite imagery: a current view,Remote Sensing and Geographial Information Systems for Resource Management in Developing Countries, pp. 71–96, 1991.Google Scholar