A cluster merging filter for noise smoothing
B.J. Liu, Keh‐Shih Chuang · IEEE Engineering in Medicine and Biology Magazine · 1994
Due to the statistical nature of X-rays and the electromagnetic field, medical images produced by these energy sources are contaminated with random noise, which degrades the image quality. Because of this effect, considerable effort has been devoted to removing noise from medical images. The authors present a new method of image noise smoothing. This method uses the natural separation of distinct populations in each window to evaluate a given pixel. In this way, more pixels that belong to the population are included in the cluster of the given pixel, and fewer pixels that do not belong to the population are erroneously included. The image smoothing is further enhanced by the join-count statistic. By implementing join-count statistics, clusters that were erroneously separated by a large variation of random noise were evaluated and merged. This operation provides better results, enhancing noise smoothing, especially in the areas with largely uniform pixels. As a result, the smoothing performance is enhanced while the preservation of edges is maintained.>