计算机工程 ›› 2008, Vol. 34 ›› Issue (4): 233-234.doi: 10.3969/j.issn.1000-3428.2008.04.083

• 人工智能及识别技术 • 上一篇    下一篇

结合有限相邻关键点的空间覆盖研究

廖明军1,3,李克平1,朱世交2,王凯英3   

  1. (1. 同济大学交通运输工程学院,上海 201804;2. 同济大学计算机科学与工程系,上海 200092;3. 北华大学交通建筑工程学院,吉林 132013)
  • 收稿日期:1900-01-01 修回日期:1900-01-01 出版日期:2008-02-20 发布日期:2008-02-20

Study of Space Covering Combined with Nearest Key Points

LIAO Ming-jun1,3, LI Ke-ping1, ZHU Shi-jiao2, WANG Kai-ying3   

  1. (1. College of Transportation Engineering, Tongji University, Shanghai 201804; 2. College of Computer Science and Engineering, Tongji University, Shanghai 200092; 3. College of Transportation and Civil Engineering, Beihua University, Jilin 132013)
  • Received:1900-01-01 Revised:1900-01-01 Online:2008-02-20 Published:2008-02-20

摘要:

从高维空间特征点覆盖的角度,定义空间点的局域特征,提出共享近邻点的覆盖算法,在空间局部区域构造各向同性覆盖,在整个区域体现了各向异性。介绍了相关共享局部覆盖算法,给出了模拟试验和标准测试集。试验结果表明,由该算法构造的分类器适合于安全性较高的应用场合。

关键词: 共享近邻点, 模式识别, 拓扑空间

Abstract: From the view point of coverage in the High-Dimensional Space(HDS), local pattern of points are defined. Covering algorithm is promoted for shared points. Anisotropy of whole space is archived by anisotropy in local space. Shared Neighbor Covering Algorithm(SNCA), analog experiment, and standard test set are defined. Experimental results show that the constructive method of SNCA is suitable to high security application.

Key words: shared neighbor point, Pattern recognition, topology space

中图分类号: