计算机工程与应用 ›› 2011, Vol. 47 ›› Issue (34): 181-183.

• 图形、图像、模式识别 • 上一篇    下一篇

一种新的类增量学习方法

秦玉平1,陈一荻1,王春立2,王秀坤3   

  1. 1.渤海大学 信息科学与工程学院,辽宁 锦州 121000
    2.大连海事大学 信息科学技术学院,辽宁 大连 116026
    3大连理工大学 计算机科学与技术学院,辽宁 大连 116024
  • 收稿日期:1900-01-01 修回日期:1900-01-01 出版日期:2011-12-01 发布日期:2011-12-01

New class incremental learning method

QIN Yuping1,CHEN Yidi1,WANG Chunli2,WANG Xiukun3   

  1. 1.College of Information Science and Technology,Bohai University,Jinzhou,Liaoning 121000,China
    2.College of Information Science and Technology,Dalian Maritime University,Dalian,Liaoning 116026,China
    3.School of Computer Science and Technology,Dalian University of Technology,Dalian,Liaoning 116024,China
  • Received:1900-01-01 Revised:1900-01-01 Online:2011-12-01 Published:2011-12-01

摘要: 提出一种新的基于超椭球的类增量学习算法。对每一类样本,在特征空间求得一个包围该类尽可能多样本的最小超椭球,使得各类样本之间通过超椭球隔开。类增量学习过程中,只对新增类样本进行训练。分类时,通过计算待分类样本是否在超椭球内判定其所属类别。实验结果证明,该方法较超球方法提高了分类精度和分类速度。

关键词: 超椭球, 类增量学习, 缩放因子

Abstract: A new Class Incremental Learning(CIL) algorithm based on hyper ellipsoidal is proposed.For every class,the smallest hyper ellipsoidal that contains most samples of the class is structured,which can divide the class samples from others.In the process of CIL,only are the samples that belong to the new incremental class trained.For the sample to be classified,its class be confirmed by the hyper ellipsoidal that it belong to.The experimental results show that the algorithm has a higher performance on classification speed and classification precision compared with hyper sphere algorithm.

Key words: hyper ellipsoidal, Class Incremental Learning(CIL), extension factor