3D腹部多排电子计算机断层扫描医学图像的非线性扩散滤波分析

Non-linear diffusion filtering analysis on 3D abdominal MDCT medical images

目的:对3D腹部多排电子计算机断层扫描(multi-detector computerized tomography,MDCT)图像进行非线性扩散滤波分析,避免低剂量MDCT图像质量改善中3D非线性扩散滤波器需要通过主观经验设置扩散参数的缺点。方法:根据3D非线性扩散滤波器的滤波原理推导扩散参数与体素值变化的依赖关系,根据该依赖关系客观地设置扩散参数值,并将加权均值策略引入非线性扩散滤波当中以抑制点噪声。结果:根据该方法设置的非线性扩散滤波器能够有效地抑制低剂量MDCT图像的量子噪声,并且能够较好地保留小病灶、血管等感兴趣区域的边缘与细节信息。结论:扩散参数与体素值变化的依赖关系可以应用在3D扩散均值滤波器的参数设置当中,使该滤波器能够有效改善低剂量MDCT图像质量。

Objective To perform non-linear diffusion filtering analysis on 3D abdominal MDCT medical images to eliminate the defects due to parameter setting based on subjective experience for low dose MDCT images. Methods The dependency between the diffusion parameter and the voxel values was deduced to set the diffusion parameters. The weighted averaging strategy was introduced to the diffusion filtering to suppress the point noise. Results The modified diffusion filtering whose parameters were set according to the deduced dependency could suppress the quantum noise effectively. On the same time, the edges and subtle information of small lesions and blood vessels, etc could be preserved well after filtering. Conclusion The dependency between the diffusion parameter and voxel values can be used to set the parameters of 3D nonlinear diffusion filter with an averaging strategy. This filter can improve the image quality of low dose MDCT images effectively.