核医学图片的数字图像处理

Digital Image Processing on Nuclear Medical Images

用线性滤波器处理核医学脏器图片．造成图像边缘不清楚，从而影响了诊断．本 文用一种非线性统计滤波器——中值滤波和平均滤波相结合来处理核医学脏器图像． 这种方法综合了平均滤波和中值滤波对噪声的抑制特性和保留图像边缘的特性．考虑 到图像的局部起伏．引入Ａｎｓｃｏｍｂｅ变换使图像的泊桑噪声与信号近似无关。最后 对实际的核医学脏器图片作了处理，并与线性滤波处理结果做了对比．收到了满意的 效果．

A nonlinear statistical filter, which mixes the median filtering with the average filtering and has the properties of the two, is used for improving nuclear medical images. The filter reduces image noise and preserves the edge sharpness as well. Because o'f the local fluctuation in the nuclear image, the Anscombe transform is utilized to separate image signals from Poisson noise. Experimental results are shown and compared with those obtained by linear filters.