基于非线性各向异性散布滤波器的磁共振颅脑图像的分割方法

A MR Image Segmentation Method Based on Nonlinear Anisotropic Diffusion Filter

目的介绍一种有效的分割磁共振颅脑图像中的大脑结构的方法。 方法它是一个多步处理过程 ,首先 ,粗略地勾画出磁共振颅脑图像中的大脑结构作为感兴趣区 ;其次 ,对感兴趣区进行各向异性散布滤波处理 ;然后 ,对滤波图像进行二值化处理作为图像模板 ;最后 ,利用该模板从原始图像中提取出大脑结构。 结果该方法能有效地分割出磁共振颅脑图像中的大脑结构。 结论利用基于非线性各向异性散布滤波器的图像处理技术能很好地从磁共振颅脑图像中分割出大脑结构。该算法具有很强的鲁棒性和稳定性。

Objective Introduce an effective method for segmenting the brain from head magnetic resonance (MR) images. A multistage process is needed. At first, roughly delineate the brain structure from head MR images as the region of interesting (ROI) manually; next, the ROI is to be processed by a nonlinear anisotropic diffusion filter; and then, the filtered image is used as an image mask after binary process; at last, the brain structure is extracted from the original image using the image mask. The brain structure of head MRI can be segmented by this method effectively. When using image processing techniques based on a nonlinear anisotropic diffusion filter to segment the brain structure in MRI, the segmentation result is correct. It has strong robustness and stability.