Generalized Multiquadric Neural Networks in Image Reconstruction

This work aims to numerically investigate the performance of multiquadric (MQ) radial basis function in more general formats for image reconstruction applications. Desired features, i.e., accuracy and shape parameter sensitivity, each form is compared explored. The famous Lena damaged using two levels damage: 20% 40%, a Salt-and-Pepper manner. It has been discovered this that β=3/2 produces reasonably good least affected by change while keeping both CPU time condition number acceptable. finding promising useful further applications MQ complex contexts.