NOISE REDUCTION IN MEDICAL IMAGES USING ADAPTIVE WEIGHTED MEDIAN FILTER BASED ON BACK PROPAGATION NEURAL NETWORK

A novel Adaptive Weighted Median Filter Based on Back Propagation Neural Network (AWM/BPNN) filter is proposed for reducing noise in medical images and improving the performance of median-based filters. The proposed filter achieves its effect through the linear combinations of the median based filters and neural network technique. In this proposed system is a three-stage process. In the first stage, the adaptive technique is used to determine whether the pixel is corrupted or uncorrupted. In the second stage, the weights of the Weighted Median (WM) filter are calculated by using Back Propagation Neural Network (BPNN) algorithm. In the final stage, the corrupted pixel value is replaced by the weighted median value. The visual and performance metrics show that the proposed filter outperforms many of the standard median filters in terms of noise removal with edge preservation.