摘要
目的:比较目测法与外标记法对SPECT/CT图像进行配准的准确性;优化外标记法配准时标记点的个数及分布。方法:利用二维运动平台系统模拟人的呼吸运动,用配准后N个外标记点在两种图像间位置平均偏差(f)作为评价指标来衡量两图像配准的精度。结果:体模静止时,不同配准方法的f值间差别不大。而在不同运动状态下对7种方法配准得出的f值进行多组间的单因素方差分析,结果表明,用≥5个标记点配准的方法均与目测法、3点配准法间差异有统计学意义,P<0.05,且前者优于后者,而用≥5个标记点配准方法差异无统计学意义,P>0.05;不同研究者间目测法得到的f值差别较大;对5点配准法进行优化表明,标记点非共面分布优于标记点共面分布。结论:目测法配准受研究者主观因素的影响较大,准确性无法保证;用非共面5个外标记点对SPECT/CT图像进行融合配准是一种准确且简单易行的方法。
OBJECTIVE： To compare the registration accuracy of SPECT/CT images using the visual method and the method of external markers, and optimize the number and distribution of markers while u sing the registration method of external markers. METHODS： By using an in house 2D motion platform system to simulate the respiratory mo- tion, the mean square deviation （f） of N external markers coordinated between two kind of images was used as a similarity evaluating index to measure the accuracy of two image registration. RESULTS： When the phantom remained static, the difference of the f values among the different registration methods was little. But under the different conditions of motion, doing single-factor analysis of variance for the f value groups of 7 registration methods, the results showed that the method using not less than five markers were obviously different from using the visual method and three marker registration method, and the former was superior to the latter（P〈0.05）, but among the registration methods using not less than five markers the differences were not obvious（P〉0.05）. For different researchers, the f values using visual method were differ ent. The optimizing results for five marker method showed that non co planar distribution of the markers was superior to coplanar distributing. CONCLUSIONS： The visual method is easily impacted by the subjective factors, and the accuracy can not be guaranteed. Using five external markers distributing non-coplanar is an accurate and simple method for SPECT/CT image registration.
出处
《中华肿瘤防治杂志》
CAS
2009年第23期1863-1866,共4页
Chinese Journal of Cancer Prevention and Treatment
基金
山东省科技攻关计划(2006GG2202012)
山东省医药卫生科研项目(HZ090)
