SPECT图像滤波函数的选择

THE SELECTION OF FILTERING FUNCTIONS ON SPECT-- BUTTERWORTH AND GAUSIAN FILTER

通过对ＳＰＥＣＴ中各种常用滤波函数的数学分析、模型试验和临床应用的比较．认为Ｂｕｔｔｅｒｗｏｒｔｈ滤波器确为ＳＰＥＣＴ最佳的重建滤波函数．经过两年来的临床证实，如果在重建前应用多功能的Ｇａｕｓｉａｎ滤波器作预处理，能进一步增强图像的对比度，提高信噪比，从而获得分辨率更高的ＳＰＥＣＴ图像．改变Ｇａｕｓｉａｎ函数的三参数Ｓ、ｆｃ和ｎ，可达到多种不同的滤波效果，因此重建前滤波首选Ｇａｕｓｉａｎ滤波器．同时对如何选择Ｇａｕｓｉａｎ滤波器的参数及临床应用的最佳条件进行了阐述．

Abstract SPECT is a very useful imaging equipment for medical diagnosis. In order to get high quality images, we must put emphasis on filter selection as well as routine quality control. Through mathematical analysis, phantom test and clinical verification, we proved Butterworth filter to be a flexible and efficient one for backprojection reconstruction. Iterative reconstruction is a much precise method but it takes much longer time to Perform. We found that another practical method to obtain higher resolution, lower noise 3-D slices if we applied Gausian filter to raw data prior to the backprojection process. Gausian filter, with three selectable parameters, could not only enhance but also filter images.