FPGA Implementation of Directional Peer-Group Image Filter

Abstract

Peer group is known as one of the simplest methodologies for image denoising in the spatial domain. Unfortunately, its filtering effectiveness will rapidly degrade while the noise density increasing. This paper introduces a modified directional peer-group filter for better restoration of images corrupted by random impulse noises. In addition, a low complexity FPGA architecture for the implementation of this simple algorithm is also demonstrated. Simulation results show that the pro-posed approach can effectively reconstruct images and preserve edges even for the image suffered from high-density noises.