本文對于由Visible Human 所提供的人體CT 圖像序列所形成的體數據場，提出了一種有效的快速中值濾波方法。中值濾波是一種非常有用的非線性濾波技術，能有效的抑制脈沖噪聲、椒鹽噪聲等。針對于醫學CT 圖像中要濾除的噪聲干擾，本文提出了一種有效的快速中值濾波方法。為了克服中值濾波的模糊邊緣特征的缺點，先人工干預在邊緣附近設置一閾值，構造一中間矩陣M，利用窗口內數據間的相關性，設計了快速的移動剔除過程，并在原排序基礎上進行快速比較排序，縮減排序過程和次數，加快了濾波速度，一定程度地避免了邊緣的模糊。
關鍵詞: 體數據；醫學圖像；中值濾波；邊緣增強
Abstract：Based on the medical slice images data set formed from Visible Human, this paper puts forward an effective and accelerated median filtering method. Median filtering is a useful nonlinear filtering technology. It can remove pulse noise and jiaoyan noise effectively. Aiming at the noises in the medical CT slice images, we put forward this effective and accelerated median filtering method in this paper. In order to overcome the disadvantage that median filtering always blurs the fringe characters of images, a threshold value is set firstly
around the edge, and thereby a midway matrix M is constructed. Meanwhile, using the pertinence of these data in this window, an accelerated eliminating and arraying process are designed. By this means, we successfully reduce the course and number of arraying and speedup the filtering and avoid blurring the edge information.
Key words: data set; medical images; median filtering; edge restoration