首页 | 本学科首页   官方微博 | 高级检索  
     

一种基于形态中值金字塔的线特征提取方法
引用本文:戴青云,余英林,张大鹏. 一种基于形态中值金字塔的线特征提取方法[J]. 华南理工大学学报(自然科学版), 2001, 29(6): 14-18
作者姓名:戴青云  余英林  张大鹏
作者单位:1. 华南理工大学电子与信息工程学院,
2. 香港理工大学计算机系,
基金项目:国家自然科学基金资助项目 (6 9772 0 2 6 ),广东省高教厅基金资助 (990 0 49)
摘    要:基于掌纹的身份识别技术是基于人体参数身份识别技术的重要内容,如何选择合适的掌纹图象及其线特征,如何提取这些线特征是该项技术的关键所在,传统的边缘检测方法易产生双边缘,。易出现位置偏移,易受噪声影响,因此不适于掌纹线特征的提取,本文提出的一种基于形态中值金字塔的多分辨率线特征提取方法,借鉴多分辨率分析的思想,结合掌纹图象的特点及其相关知识来提取线特征,取得了较理想的实验结果。

关 键 词:掌纹图象 线特征提取 多分辨率分析 形态中值金字塔
文章编号:1000-565(2001)06-0014-05

A Line Feature Extraction Method Based on Morphological Median Pyramid
Dai Qing_yun Yu Ying_lin David Zhang. A Line Feature Extraction Method Based on Morphological Median Pyramid[J]. Journal of South China University of Technology(Natural Science Edition), 2001, 29(6): 14-18
Authors:Dai Qing_yun Yu Ying_lin David Zhang
Affiliation:Dai Qing_yun 1 Yu Ying_lin 1 David Zhang 2
Abstract:The personal identification technology based on palmprint is an important means to automated biometrics technologies and systems, in which how to select the proper line features and how to extract these features are the most challenging issues. There are three kinds of lines and curves in palmprint images, and these line features are the most important clues for identifying or verifying a person's identity. The traditional edge detection methods are not appropriate to palmprint images because of their shortcomings such as double_line edges, location offsets, sensitivity to noise and so on. The line feature extraction method based on morphological median pyramid presented here is well supported by experimental results.
Keywords:palmprint image  line feature extraction  non_linear multi_resolution analysis  morphological median pyramid
本文献已被 CNKI 维普 万方数据 等数据库收录!
设为首页 | 免责声明 | 关于勤云 | 加入收藏

Copyright©北京勤云科技发展有限公司  京ICP备09084417号