1. 青岛大学自动化工程学院,山东,青岛,266071
2. 中国科学院半导体研究所神经网络研究室,北京,100083
3. 青岛大学自动化工程学院山东青岛,266071
4. 中国科学院半导体研究所神经网络研究室北京,100083
网络出版:2008-12-25,
纸质出版:2008
移动端阅览
杨国为, 王守觉, 刘扬阳. 仿生模式识别的两个关键技术问题研究[J]. 电子学报, 2008,36(12):2490-2492.
YANG Guo-wei, WANG Shou-jue, LIU Yang-yang. Research on Two Key Technical Problems in Biomimetics Pattern Recognition[J]. Acta Electronica Sinica, 2008, 36(12): 2490-2492.
同类事物特征在高维空间中形成的复杂几何形体的局部连续(连通)方向和通路的确定问题
以及找到满意的该复杂几何形体的覆盖体
判断高维空间一点是否在该覆盖体内的问题
是在传统的数学分析框架内的不适定难题.本文用高维空间点分布分析方法给出了解决以上两个仿生模式识别的关键技术问题的有效方法.本技术方法能将同类相近样本归于一个连通几何形体内
能把不同类的相近样本归于不同几何形体内
从而实现具体事物的精准识别.
The problem of determining the local continuity(connection)direction and of the route of the complicated geometrical body formed by characteristics of the same type things in high dimension space
and the problem of finding a satisfactory cover for this complicated geometrical body and deciding whether or not a high dimension point is within the cover are problems difficult to solve within the framework of the traditional mathematical analysis.This paper provides an effective method for the solution to the key technical problems of the above two biomimetics pattern recognition by use of high dimension point distribution analysis method.This technical method is capable of grouping similar samples of the same type in one connected geometrical body
and similar samples of different types in different geometrical bodies
thus achieving the precise recognition of concrete things.
0
浏览量
829
下载量
6
CSCD
关联资源
相关文章
相关作者
相关机构
京公网安备11010802024621