引用本文: | 温涛,蔡宣平.一种基于置信度的动态特征提取算法[J].电力自动化设备,2001,(3):23-25,48 |
| .Dynamic Algorithm of Feature Extraction Based on Believing-Degree[J].Electric Power Automation Equipment,2001,(3):23-25,48 |
|
|
摘要: |
结合实际工程中的某些要求,将统计学中的置信度概念引入和持征提取中,提出了一种按照精度要求进行有效的特征提取的算法,该算法能够实现在抑制漏警错误或者虚警错误的同时降低特征维数的功能,最后介绍了此算法成功地应用于现钞识别于现钞识别系统的系统设计细节以及实验结果,从实际应用于现钞识别系统的系统设计细节以及实验结果,从实际应用上证明了该算法的有效性。 |
关键词: 置信度 模式识别 动态特征提取算法 统计学 |
DOI: |
分类号:TP391.4 |
基金项目: |
|
Dynamic Algorithm of Feature Extraction Based on Believing-Degree |
WEN Tao CAI Xuan ping JIN Ou
|
Abstract: |
Combined with the requirement of real project,the statistic concept of believing degree is introduced into feature extraction and a dynamic algorithm for effective feature extraction is proposed,which is based on accuracy requirement and capable of constringing the feature dimension while reducing leaked or pretend alarm probability.The design detail of a note identification system,which adopts the algorithm,and its test results are introduced.The real applications have verified its effectiveness. |
Key words: feature extraction,statistics,believing degree, |