An Improved Edge Preserving Smoothing Method (IEPS)

Edge preserving smoothing prior to image segmentation is very important, since the image detailed information is mostly in edges, and smoothing uncorrupted pixels can sharply reduce the image quality. Hence, a novel noise detection method is essential to identify the noised pixels and noise-free pixels. In this paper, the noise detection is based on the difference between the pixel of interest value and its eight neighbors. The proposed method uses the mean of the five pixels with minimum difference to distinguish edge pixels from noisy pixels. Then, it is combined with the edge preserving smoothing using nine rotating mask to replace the noised pixel. The results show that the proposed method outperforms other method, with higher Peak Signal to Noise Ratio and reducing noise.