A New Hybrid Filtering Technique Based on Neighboring Pixels to Remove Impulse Noise from Digital Images

Abstract

This paper proposes a hybrid technique to remove impulse noise from digital images. In this approach the filtering operation is based on 33 neighborhood of pixel under consideration. During filtering, the properties of neighborhood are considered to check whether it is highly corrupted with noise, medium or only itself act as impulse. Based upon these properties a new hybrid technique has been proposed to process the pixel which further uses different schemes. The experiments have been performed at various noise levels on standard images as well as on real images. The results have been evaluated on the basis of metrics like Signal to noise ratio (SNR), Edge preservation index ( EPI), Structure similarity index measure (SSIM), Multi scale structure similarity index measure (MS-SSIM) and Peak signal to noise ratio ( PSNR). From the results, it has been observed that proposed technique has worked efficiently by preserving the edges and fine lines. To demonstrate the effectiveness of proposed technique, the results have also been compared with other well accepted denoising techniques