Image noise removal on improvement adaptive medium filter
Di Jia Wu · Jiguang zazhi · 2009
Traditional adaptive median filtering performs badly on removing noise and takes a long run time especially when there's a big difference between the stable maximum and minimum window size.The algorithm that we have proposed,first estimates the thickness of asingle image's salt-and-pepper noise based on the noise's distribution,and it found a functional relation between thickness and window size.After the two window sizes are confirmed by the noise thickness,the algorithm does median filtering on the image.Our experiment shows that the algorithm changes run time according to the change in the noise thickness.With changes in PSNR,we can see that the algorithm performs well to remove the noise.