A Simple and Efficient Median Filter for Removing High-Density Impulse Noise in Images

Abstract

This paper presents a simple median filter, which first applies the adaptive median filter and then evaluates a median value to perform an algorithm. Conventionally, most of the median filters have excellent performances only for the corrupted images with low-density impulse noise, but they usually do not perform very well for high-density impulse noise. Although some filters with a noise detector perform well for high-density noise, their algorithms are usually complicated and time consuming. Besides, their performances on low-density noise have not been presented yet. The proposed median filter can efficiently remove both low- and high-density impulse noises in corrupted images, and the experiments show its filtering performances by comparing with some conventional median filters.