基于张量形态学梯度和图像森林化变换分水岭算法的扩散张量图像脑白质分割

Segmentation of Diffusion Tensor Image of Brain White Matter Tissues Based on Tensorial Morphological Gradient and Image Forest Transformation Watershed Algorithm

目的基于张量形态学梯度和图像森林化变换分水岭算法,提出一种从脑扩散张量图像中分割出白质组织的方法,并验证其是否快速、有效。方法首先总结了几种张量相似性函数,进一步计算张量形态梯度,再利用基于图像森林化变换分水岭算法对脑扩散张量图像进行白质组织的分割。结果采用此方法对志愿者的脑扩散张量图像进行分割,得到了较好结果。结论该方法易于实现且鲁棒性强,并适合进一步对脑白质亚结构体的分割。

Objective To put forward a new method for getting the segmentation of white matter（WM） from brain diffusion tensor image（DTI）,and verify if it was quick and effective.Methods Firstly several tensorial similarity functions were enumerated,then tensorial morphological gradient（TMG） images were computed.Finaly the segmentation of the TMG images were accomplished based on image forest transformation（IFT） watershed algorithm.Results A patient＇s brain diffusion tensor image dataset was used to test this method,and a satisfactory result was obtained.Conclusion The proposed method is robust and can be implemented simply.It can be further applied to the segmentation of sub-construction of brain white matter.