Random-valued impulse noise removal from images: K-means and luo-statistics based detector and nonlocal means based estimator

Impulse noise detection is the key issue, while removing random-valued impulse noise from digital images. In this paper, we present a new impulse detection algorithm based on combination of Luo-statistic and k-means clustering. This paper also presents a novel approach to measure impulse noise density level in the corrupted image, knowledge of which allows us to select suitable parameters for the noise detection algorithm. In noise filtering stage, we apply nonlocal-means (NL-means) estimator to restore noisy pixels to their actual values. Extensive experimental results show that the proposed method outperforms most of the existing impulse noise removal techniques both in terms of noise detection and image restoration.