基于Hessian矩阵跟踪算法在血管新生量化中应用.docVIP

基于Hessian矩阵跟踪算法在血管新生量化中应用 　　摘 要：血管新生与肿瘤的生长密切相关，CAM（鸡胚绒毛尿囊膜）是研究血管新生的常用载体，所以通常对CAM中特定区域的血管分割量化来衡量抗肿瘤药物对血管的抑制作用。本文提出一种CAM图像分割算法，首先利用Hessian矩阵下多尺度分析得到一个初始的CAM血管轮廓，在此基础上跟踪未知血管，得到的最终分割结果。最后，对分割出的CAM血管进行骨架提取、裁剪等处理，计算出图像中血管的密度、长度和分叉点个数等量化指标，综合评价血管的生长情况。在相同系列的CAM图像上和其他方法进行实验对比。结果显示基于Hessian矩阵跟踪算法可以为血管新生状况提供有效的分析，尤其对微小血管的检测比较优越，而且具有更高的准确率。 　　关键词：血管新生；Hessian矩阵；鸡胚绒毛尿囊膜图像；跟踪算法 　　中图分类号：TP391.41 文献标识号：A 文章编号：2095-2163（2015）01- 　　Abstract： Angiogenesis has a close relationship with tumor， one of the most popular assay tissues for the research of angiogenesis phenomenon is the （CAM） chick embryo and its chorioallantoic membrane. So the segmentation and quantitative of a specific area of CAM blood vessels is usually used to measure the inhibitory effect of antineoplastic drugs on blood vessels. The paper proposes a segmentation algorithm of CAM vessel. First， the paper gets the coarse CAM vessel segmentation results through the multi-scale of Hessian Matrix. Then the unrecognized blood of CAM vessels are extracted by tracking those initial results. After that， the vessel segmentation result is made out under the overall consideration of different scales. Finally， a lot of steps to process the vascellum for quantization have been implemented in detail， which mainly includes blood vessel skeleton， pruning and so on. Based on the aboved， according to the index of vascellum， such as density， length， and branch point and so on， the evaluation on the growth of CAM vascellum can be made comprehensively. Through experiment with other methods on the same CAM images， results show that vessel segmentation based on Hessian matrix and tracking can provide effective and accurate analysis for the angiogenesis， especially to the tiny CAM vascellum. 　　Keywords： Angiogenesis； Hessian Matrix； CAM Image； Tracking Algorithm 　　0 引 言 　　血管新生[1]是一种常见的生理现象，可为肿瘤细胞的生长提供所必需的营养和氧气，因此血管新生是肿瘤生长的一个必要条件。目前，研究血管新生现象最常见的一种实验载体是发育中的鸡胚绒毛尿囊膜（CAM）[2-4]。可以通过对CAM中特定区域的血管分割量化来衡量血管新生药物对血管的促进或者抑制作用。 　　1 CAM图像分割 　　CAM