A robust noise removal algorithm with consideration of contextual information

Abstract

This paper analyzes an image noise model of additive positive and negative impulses that often appear in practical applications. Based on the characteristic that any pixel in an undisturbed image is similar to its neighbors, a local pixel correlation coefficient is proposed. For a pixel, based on the number of similar pixels in its neighborhood, the probability of whether it is noisy or normal can be accurately calculated. An adaptive masking weighted mean filter with consideration of contextual information is proposed to filter noise while retaining the edge details of the image. The proposed algorithm does not require any initial parameters or threshold values to be set. Experimental results show that the proposed algorithm is applicable to the proposed noise model and that the proposed noise filtering is significantly better than that of existing algorithms.

Effective and adaptive algorithm for pepper‐and‐salt noise removal

“…According to the noise model expressed by (2), for pixel f ( i , j ), the grey level is in [ m , 255 − m ], which is a pixel with an unpolluted signal; otherwise, it is a pixel with suspected noise. Pixels in an unpolluted image often have a grey level similar to their neighbourhood [20]. Therefore, for a pixel with suspected noise, if the number of similar noise‐free pixels in its neighbourhood or the number of pixels with closed grey level in its neighbourhood is sufficient, the pixel would be regarded as a noise‐free pixel; otherwise, it is a noisy pixel.…”

“…The setting of δ is obtained from the experimental result based on some estimation according to Chen and Chang [20]. δ 1 , δ 2 and δ 3 are used to be three thresholds of N sp1 ( i , j ), N sp2 ( i , j ) and N sp1 ( i , j ) + N sp1 ( i , j ) for L = 1. δ 4 , δ 5 and δ 6 are for L = 2.…”