定量描述颅面X光片图像的相似性的一种方法

目的:提出一种基于特征点位置坐标的定量描述颅面x光片图像的相似性的方法,并应用到研究颅面X光片图像正常值的确定中.方法:提取两幅X线头影测量图像中的特征点的位置坐标,构成两组二维向量,用典型相关分析这两组向量的相关性,该相关性就表征了两幅图像间的相似度.根据图像的相似性,获得各个图像叠加时的权重,从而确定正常值.结果:用该方法分析10例男性恒牙期正常颅面X光片,得到每两幅图之间的相似度和各图叠加时的加权系数.并对影响相似度的因素作了分析,说明特征点的数目和选取方式对相似度有影响,相似度越低,越容易受影响.重复定点中的随机误差对相似度的影响很小.结论:该方法能够定量地描述图像的相似度,对研究颌面正常值具有一定的意义,并可以直接推广到研究3D医学图像的相似度和正常值.

A METHOD TO EVALUATE THE SIMILARITY BETWEEN THE CEPHALOMETRIC RADIOGRAPHS QUANTITATIVELY

Objective:The purpose of this paper is to pose a method that analyzes the similarity between the cephalometric radiographs based on the landmark's positions, and applies it to form the norms for the cephalometric radiographs analysis Methods: The x, y coordinates of the landmarks was got to make up two groups of vectors, and the correlation between them,which represents the similarity of the two images was analyzed The norms were obtained by calculating the weight based on the similarity Ten cephalometric radiographs images were analyzed in this method and weight of every image was calculated to gain the norms, and the factors that affect the similarity were discussed Results: How many landmarks were used and which ones were chosen influenced the results remarkably, the lower the similarity was, the more the influence was But random errors had little influence on the similarity Conclusion:This method can evaluate the similarity between the images quantitatively and can also be applied to the 3D images