Numerical Implementation of the Hopfield-Type Neural Networks from the MEVA Method in Remotely Sensed Images

Abstract

In this paper we present a new outlook of the numerical approximation for implementing of the Hopfield-type Neural Networks (HNN) to the computational processing of Remotely Sensed Images (RSI). Here, we implemented the fused Maximum Entropy Variational Analysis (MEVA) method that presents the distinguished reconstruction strategy for image enhancing just by one process. The numerical implementation is based on the Jacobi and Gauss-Jordan Methods for solving the energy minimization problem. Therefore, we present several selected computer simulation examples where real images as addressed to illustrate the outstanding usefulness of this method. Likewise, we present some quantitative and qualitative analysis to the improvement of the new approximation of the MEVA method