Adaptive region growing impulse noise estimator for color images

摘要

In this paper, a novel region growing impulse noise estimator for color images is proposed. The aim of this estimator is to distinguish noisy pixels from uncorrupted pixels and subsequently measure the noise proportion efficiently. We use a region growing technique to segment the images into clusters of pixels and propose an adaptive decision scheme to measure the noise proportion. Performance analyses show the proposed scheme outperforms some of the state-of-the art techniques.