基于B样条混合模型的一维形状分析  

One-dimensional Shape Analysis Based on B-spline Mixture Model

作　　者：邵周健 王旭辉[1] SHAO Zhoujian;WANG Xuhui(School of Mathematics,Hefei University of Technology,Hefei 230601,China)

机构地区：[1]合肥工业大学数学学院,合肥230601

出　　处：《大学数学》2022年第4期14-20,共7页College Mathematics

基　　金：国家自然科学基金(61772167)。

摘　　要：高斯混合模型被广泛应用于基于地标的形状表示与分析,然而该模型在表示形状时,其对应的信息矩阵通常不具有显式表达式,导致计算相对复杂.为了提高计算效率,给出B样条混合模型用于表示形状.由于B样条函数的局部支撑性,使得基于B样条混合模型的形状分析计算相对高效.通过数值实例说明B样条混合模型在形状分析上的有效性.Gaussian mixture model is widely used in shape representation based on landmarks.However,the corresponding information matrix of model usually does not have explicit expression in the shape representation,which lea d s to high computational cost.In order to improve computational efficiency,the B-spline mixed model is used to represent shapes.Due to the local supported property of B-spline function,the calculation with shape analysis based on B-spline mixture model is efficient.Numerical examples are provided to illustrate the effectiveness of B-spline model in shape analysis.

关 键 词：B样条模型 测地线 黎曼流形 熵度量 

分 类 号：O241[理学—计算数学]