基于特征点扩充及PCA特征提取的ASM定位算法  

ASM Location Algorithm Based on Feature Points Expansion and Features Extracted by PCA

作　　者：罗声平 姚剑敏[1] 郭太良[1] 

机构地区：[1]福州大学物理与信息工程学院,福州350002

出　　处：《光电子技术》2015年第1期39-44,48,共7页Optoelectronic Technology

基　　金：国家863重大专项(2012AA03A301;2013AA030601);国家自然科学基金(61101169;61106053);福建省自然科学基金(2011J01347)项目资助

摘　　要：活动形状模型ASM(Active Shape Model)在目标对象的定位中得到广泛应用,但传统ASM算法定位精度较低,模型容易收敛到错误位置,为此,本文提出了一种基于特征点扩充和主成分分析PCA(Principal Component Analysis)灰度特征提取的ASM改进算法:首先,采用等距插值的方法扩充手工标定的特征点;其次,提出采用主成分分析PCA处理特征点法线灰度信息代替原算法中的灰度值求导,统计特征点局部灰度特征,以提高目标定位的精度。实验结果表明,与传统ASM算法相比,本文的改进算法的目标定位精度和鲁棒性都有了显著的提高,实验数据显示,平均定位误差降低了38%以上。ASM（Active Shape Model）algorithm has been being widely used in location of the target object.However,the localization accuracy of the traditional ASM algorithm is low and the model tends to converge to a wrong location easily.So an improved ASM algorithm based on feature points expansion and gray features extracted by PCA（Principal Component Analysis）is proposed. In order to improve the accuracy of ASM algorithm in target location,equidistant interpolation is applied to the expansion of feature points firstly;Secondly,PCA is applied to the processing of the normal gray information instead of the derivation of grey value.It is experimentally indicated that the a significant increase in the localization accuracy and robustness is realized with the improved algorithm with the average localization error decreased by more than 38%.

关 键 词：特征点扩充 主成分分析 配准变换 匹配收敛 

分 类 号：TP391.4[自动化与计算机技术—计算机应用技术]