In this paper a novel approach for de noising images corrupted by randomvalued impulses has been proposed. Noise suppression is done in two steps. Thedetection of noisy pixels is done using all neighbor directional weightedpixels (ANDWP) in the 5 x 5 window. The filtering scheme is based on minimumvariance of the four directional pixels. In this approach, relatively recentcategory of stochastic global optimization technique i.e., particle swarmoptimization (PSO) has also been used for searching the parameters of detectionand filtering operators required for optimal performance. Results obtainedshows better de noising and preservation of fine details for highly corruptedimages.
展开▼
机译：本文提出了一种新颖的，用于通过随机脉冲脉冲损坏的发出图像的新方法。噪音抑制是以两步完成的。使用5 x 5窗口中的所有邻维定向权重池（ANDWP）完成嘈杂像素的注意。滤波方案基于四个方向像素的最小性。在这种方法中，对于随机全局优化技术的相对近来，粒子群化（PSO）也已被用于搜索最佳性能所需的检测和滤波器的参数。结果获得了更好的腐败细节的更好的通知和保存。
展开▼
