DIBS: distance- and intensity-based separation filter for high-density impulse noise removal

Noise is an unwanted element that degrades the quality of digital images. Salt and pepper noise is a type of noise that is introduced in one or more steps during image acquisition, enrolment, or transmission. It is therefore important to apply superior restoration methods to mitigate the noise. In this paper, a novel distance- and intensity-based separation filter is proposed wherein the denoised image is obtained by processing subsets of noise-free pixels. The core concept revolves around discarding less relevant information to get a smaller set of relevant pixel values. This filter removes color streaks and distortions that often appear in other filters at high salt and pepper noise. The quantitative comparisons on various standard images reveal that the proposed method outperforms state-of-the-art noise removal filters in terms of overall image detail restoration; achieving better noise removal, especially at higher noise levels. A high-level hardware architecture is also provided.