Developing a Neural Network to Act as a Noise Filter

摘要

This study uses the neural network simulator called NETS to determine if neuralnetworks could perform a non-linear filtering operation to remove noise from two-dimensional (2-D) data and produce a noise-free image. Application is geared toward the development and performance of neural network filters, including the development of an optional neural network architecture and the use of criteria in determining how accurate the net filtered noise to produce a noise-free image.