Computer Engineering and Applications ›› 2007, Vol. 43 ›› Issue (30): 228-230.

• 工程与应用 • Previous Articles     Next Articles

Degree of confidence-based handwritten numeric recognition using combining classifiers for complementary

WU Jian-hui1,2,ZHANG Guo-yun1,YANG Kun-tao2   

  1. 1.Dept.of Physics & Electronics Information,Hunan Institute of Science & Technology,Yueyang,Hunan 414006,China
    2.School of Optoelectronics Science & Engineering,Huazhong University of Science & Technology,Wuhan 430074,China
  • Received:1900-01-01 Revised:1900-01-01 Online:2007-10-21 Published:2007-10-21
  • Contact: WU Jian-hui



  1. 1.湖南理工学院 物理与电子信息系,湖南 岳阳 414006
    2.华中科技大学 光电子科学与工程学院,武汉 430074
  • 通讯作者: 吴健辉

Abstract: Off-line handwritten numeric recognition is an important branch of Optical Character Recognition(OCR).It is widely used in the fields of bank note and post code handwritten numeric recognition task.There are many difficult for improve recognition rate if use only one classify machine,so many combining classifies methods have been proposed.In this paper,through extracting the feature of the off-line handwritten numeric characters based on the feature complementary,a method of integrated multi-classifiers of complementary based on the degree of confidence is proposed,witch includes the technology of the least distance classifier,the tree classifier as well as the BP network classifier.Experimental results show that the method proposed in this paper has excellent recognition rate and speed.

摘要: 离线手写数字识别是光学字符识别的一个重要分支,在银行票据识别、邮政编码识别等领域有着广泛的应用。由于单一分类器在识别率上很难达到要求,人们提出了各种集成分类器识别方案。通过对离线手写数字的特征提取,从特征互补的角度出发,采用了最小距离分类器、树分类器和BP网络分类器进行多分类器互补集成,提出了基于置信度的多分类器互补集成方法。通过实验对比,基于置信度的多分类器互补集成手写数字识别在识别率和识别速度上达到了满意的结果。