摘要
针对现有的血管分割方法会因图像中噪声的影响,而不能将断裂的血管部分分割出来,导致分割精细程度不够高的问题,提出一种新的基于八元数Cauchy积分公式的血管分割新算法。该方法首先根据血管走向构造合适的六维向量,提取出体数据中的每一个像素点六维方向上6个邻域点的像素值,并将其构造成一个八元函数。然后,构造一个封闭的立方体区域,将封闭立方体区域边界上26个点对应的八元函数的平均作为中心点的八元函数,并将其表示成种子点特征向量。最后,遍历图像数据,用种子点的特征向量与每一点的特征向量做乘积,通过判断内积值是否接近1,且叉乘积的模是否接近0,来判断该点是否为血管中的点。实验结果表明,对肝脏血管的分割,该算法能够快速有效地分割出更多更精细的血管。八元数这一高维的数学工具,在计算过程中综合考虑多个特性,与少量特性进行处理的结果相比较,有着明显的优势,相对于传统的血管分割算法,提高了血管分割的精度。
Because of the influence of the noise in the image,the cracked blood vessel can not be separated,which leads to the problem that the segmentation precision is not high enough. This paper proposes a new three-dimensional region growing algorithm based on high dimensional mathematical tools: octonion vector product. The proposed method first uses traditional region growing algorithm to segment continuous blood vessels. The edges of the vessel are found by gradient calculation and they are used to be initial seeds in the octonion growing regions. Then the six neighborhood of seeds' gray are used to construct octionions which represent the characteristics of seeds. The octonion vector product operation is used to connect the rupture vasculars which are under the influence of noise. Experimental results of the liver vessel segmentation show that the algorithm can quickly and efficiently segment more vessels. This octonion mathematical tools of high dimensions,which can consider multiple characteristics in the process of calculation,and this method has obvious advantages compared with the traditional vascular segmentation algorithm,and improved the accuracy of blood vessel segmentation.
出处
《微型机与应用》
2017年第17期45-48,55,共5页
Microcomputer & Its Applications
基金
国家高技术研究发展计划(863计划)(2012AA021105)