Novel peer group filtering method based on the CIELab color space for impulse noise reduction

This paper presents a novel peer group filtering method, called the NPGF method, for impulse noise reduction. The main contributions of the proposed method are two-fold. First, we propose that the impulse noise detection is performed in the CIELab, instead of the RGB, color space to enhance noise detectability. Secondly, the proposed method employs two different-sized windows in determining the status for each pixel, alleviating the problems in correcting non-corrupted pixels in the neighborhood of edges in the textured regions. Based on three typical color images, experimental results demonstrate that the proposed NPGF method achieves better performance in noise detection when compared to the existing method.