The LSF restoration problem is written as a Maximum Entropy one, where the
constraint on the restoration energy is dictated by the "Discrepancy Princi
ple". The ME solution is found by means of a continuous-Hopfield neural net
work which reduces the energy of the output misfit, and maximizes the resto
ration entropy at the same time. A positive learning parameter controls the
constraint compliance. Prior knowledge insertion into the net's algorithm,
such as prior LSF models, upper bounds, etc. is presented. Simulations, bo
th with computer generated and experimental data are carried out. The resul
ts are compared to those of the Least Squares method. Sensitivity of constr
aint fulfillment is analyzed. (C) 1999 Elsevier Science B.V. All rights res
erved.