当前位置:主页 > 科技论文 > 软件论文 >

基于特定方向多尺度稀疏编码的手掌静脉特征识别方法

发布时间:2018-05-08 08:42

  本文选题:稀疏编码 + 掌静脉识别 ; 参考:《北京邮电大学学报》2017年05期


【摘要】:为提高手掌静脉特征识别的准确率和效率,提出了基于特定方向多尺度稀疏编码(DMSC)方法.首先计算全局方向和局部方向,然后利用多尺度Gabor对划分的子区域进行独立的多尺度方向滤波,并使用稀疏编码方法进行特征提取,利用外形编码产生特征矩阵,使用标准化汉明距离方法对特征矩阵进行相似性判断.实验结果表明,DMSC方法的类间区分性较好,识别力和效率高于现有的手掌静脉识别方法.
[Abstract]:In order to improve the accuracy and efficiency of palmar vein feature recognition, a multi-scale sparse coding (DMSC-based) method based on specific directions was proposed. First, the global direction and local direction are calculated, then the multi-scale Gabor is used to filter the sub-regions, and the sparse coding method is used to extract the feature, and the contour coding is used to generate the feature matrix. The method of normalized hamming distance is used to judge the similarity of feature matrix. The experimental results show that the DMSC method has better discriminating ability and efficiency than the existing palmar vein recognition method.
【作者单位】: 北京邮电大学信息与通信工程学院;
【分类号】:TP391.41


本文编号:1860724

资料下载
论文发表

本文链接:https://www.wllwen.com/kejilunwen/ruanjiangongchenglunwen/1860724.html


Copyright(c)文论论文网All Rights Reserved | 网站地图 |

版权申明:资料由用户c2a0a***提供,本站仅收录摘要或目录,作者需要删除请E-mail邮箱bigeng88@qq.com