基于快速匹配算法的颈动脉超声图像的运动分析  

The Motion Tracking of Carotid Ultrasound Image based on Fast Block Matching Algorithm

作　　者：徐明才[1] 陶进绪[1] 连娟[1] 李昕[1] 吴学咏[1] 汪华[1] 

机构地区：[1]中国科学技术大学电子工程与信息科学系,合肥230027

出　　处：《生物医学工程研究》2009年第3期175-179,共5页Journal Of Biomedical Engineering Research

摘　　要：颈动脉超声图像中的运动信息能够间接地反应颈动脉弹性等状况,结合颈动脉内中膜厚度(CIMT)能够为心脑血管疾病诊断提供定性与定量的依据。我们将改进后的金字塔快速匹配算法(modifiedblock sum pyramid,MBSP)应用于颈动脉超声波图像斑点跟踪获得运动信息。理论和实验结果都表明,改进后的金字塔块匹配算法能有效地减少运动跟踪的运算量,并且有着和改进前的金字塔块匹配算法相同的准确度。运动跟踪结果能够为医生诊断心脑血管疾病起到一定的辅助作用。The motion information in the carotid ultrasound image can reflect the condition of carotid whether it has hardened, expanded and contracted in the right way. Then it can help doctor to diagnose cardio cerebral vascular diseases together with the carotid intima - medial thickness（CIMT） .We used the modified block sum pyramid （MBSP） algorithm to analyze the motion of carotid ultrasound images. The result shows that the modified algorithm effectively reduced the computation while speckle tracking. Besides that, the MBSP algorithm has the same accuracy with the block sum pyramid. The tracking result can assist the doctor to evaluate the disk of eardio cerebral vascular diseases to a certain degree.

关 键 词：颈动脉 斑点跟踪 金字塔算法 运动分析 平均绝对差函数 

分 类 号：R318[医药卫生—生物医学工程]