基于图像分析的鼻纹特征点对比分析方法研究  

Contrast Analysis of Feature Points for Nasal Lines Based on Image Processing

作　　者：胡郁 

机构地区：[1]上海出版传媒研究院,上海200093

出　　处：《计算机仿真》2016年第3期314-317,共4页Computer Simulation

基　　金：教育部高校硕士点基金(200801120007)

摘　　要：由于鼻纹特征点发生变化时,无法提取细致的鼻纹特征,采用传统的算法进行鼻纹特征点对比分析中,缺少准确的鼻纹特征点,导致分析精确度低、适应性差的问题。提出采用KPCA-KFDA融合算法的鼻纹特征点对比分析方法,首先将核主分量分析(KPCA)和核Fisher判决分析(KFDA)提取的鼻纹图像的特征点进行组合,然后建立高质量的鼻纹图像的特征点向量,融合于支持向量机的分类策略,依据较少鼻纹图像的特征点训练样本对鼻纹图像的特征点进行精确的分类识别,最后精确地完成鼻纹特征点对比分析。仿真结果证明,KPCA-KFDA融合算法的鼻纹特征点对比分析方法分析精确度高,适应性强。A contrast analysis method of feature points for nasal lines based on fusion algorithm of the kernel prin- cipal component analysis （KPCA） and the kernel Fisher decision analysis （KFDA） is presented. This method com- bines KPCA with KFDA to extract feature points of nasal lines, the feature point vector of high quality image of nasal lines is established, and it is fused in the classification strategy of support vector machine （SVM）. According to the training samples of feature points for nasal lines image, the feature points of nasal lines image are accurately classified and identified, then the contrast analysis of feature points for nasal lines is accurately finished. Simulation proves that the proposed method has high precision and strong adaptability.

关 键 词：鼻纹 特征对比 核主分量分析 

分 类 号：TP391[自动化与计算机技术—计算机应用技术]