当前位置: X-MOL 学术Multimed. Tools Appl. › 论文详情
Our official English website, www.x-mol.net, welcomes your feedback! (Note: you will need to create a separate account there.)
Four stage median-average filter for healing high density salt and pepper noise corrupted images
Multimedia Tools and Applications ( IF 3.6 ) Pub Date : 2020-08-26 , DOI: 10.1007/s11042-020-09557-3
Bharat Garg , K. V. Arya

This paper introduces a novel four stage filter algorithm to ameliorate images corrupted by very high density salt-and-pepper noise. The proposed algorithm exhibits two parallel trimmed median filters (TMF) at the initial stage followed by a masking logic that selects denoised pixel based on the priority. To reduce the blurring effect, higher priority is given to TMF with small window size. In the absence of noise-free pixels, the current extreme pixel is left unchanged at the first stage. Further, the denoising of unprocessed extreme pixels is done with TMF of large size window at the second stage. The remaining noisy pixels are improved by the running average filter at the third stage. Finally, the last stage handles the noisy pixels at the boundary and rare scenario. Since the proposed filter utilized non-extreme pixels to estimate denoinsed pixels value, it effectively eliminates salt and pepper noise along with better edge preservation. The performance analysis of the proposed filter is carried out with various benchmark images for varying noise density. The experimental results demonstrate on an average improvement of 2.09 dB (0.018) and 1.06 dB (0.0478) of PSNR (SSIM) respectively for wide (10% - 90%) and very-high (90% - 98%) noise density ranges over state-of-the-art filters.



中文翻译:

四级中值平均滤波器,用于修复高密度盐和胡椒噪声损坏的图像

本文介绍了一种新颖的四级滤波算法,以改善由于极高密度的椒盐噪声而损坏的图像。所提出的算法在初始阶段展示了两个并行修整的中值滤波器(TMF),然后是根据优先级选择去噪像素的掩蔽逻辑。为了减少模糊效果,对具有较小窗口大小的TMF给予更高的优先级。在没有无噪声像素的情况下,当前的极限像素在第一阶段保持不变。此外,在第二阶段使用大尺寸窗口的TMF对未处理的极端像素进行去噪。在第三阶段,通过运行平均滤波器可以改善剩余的噪点像素。最后,最后一步处理边界处和稀有场景中的嘈杂像素。由于所提出的滤波器利用非极端像素来估计去像素像素值,因此它可以有效消除盐和胡椒噪声,并具有更好的边缘保留性。针对各种噪声密度,使用各种基准图像对所提出的滤波器进行了性能分析。实验结果表明,在整个(10%-90%)和非常高(90%-98%)的噪声密度范围内,PSNR(SSIM)分别平均提高了2.09 dB(0.018)和1.06 dB(0.0478)最先进的过滤器。

更新日期:2020-10-17
down
wechat
bug