摘要
传统模式识别是一种"微观宏观"的方法,模式拒识或误识的根源在于特征提取和选择的不当。本文提出一种"宏观微观"的分布式模式识别框架方法,较之传统方法:(1)回避了模式识别较难处理的特征抽取和选择问题,为复杂对象识别提供新的方法和技术;(2)采用动态、分布计算环境取代传统的静态、集中式模式识别环境。
The methodology of tradition pattern recognition is a that of from macroscopic to microcosmos,the source of a pattern is refused or mistake recognition lie in impropriety abstraction and choicing the character. A framework of distributed pattern recognition be presented in this paper, it is a methodology of from microcosmos to macroscopic. The main innovation are:(1) advoid the difficulty of abstraction and choicing the character, provide a new technology for complex object recognition;(2) spread pattern recognition of static sate and concentration into dynamic state and distributed.
作者
蔡艳婧
王则林
CAI Yan-jing;WANG Ze-lin(Electronics and IT Department/Jiangsu Vocational College of Business,Nantong 226011,China;School of Information Science and Technology/Nantong University,Nantong 226000,China)
出处
《山东农业大学学报（自然科学版）》
北大核心
2020年第2期266-268,共3页
Journal of Shandong Agricultural University：Natural Science Edition
基金
江苏省高校高端研修资助项目(2018GRFX022)
江苏高校“青蓝工程”项目资助(苏教师[2019]3号)。
关键词
分布式
模式识别
Distribution
pattern recognition
