Off-line Handwritten Numeral Recognition Based on PSO-BP Neural Network

摘要：

The handwritten form numeral off-line recognition always is hot topic in the pattern recognition research. The handwritten form numeral has variational and capriciousness characteristic. High distinguishes rate is requested in the practical application. In view of this question, off-line recognition sorter is designed based on the POS-BP neural network handwritten form numeral and deferred to the standard handwritten form standard to design the training sample and the test sample. The recognition rate has achieved 83％ through the simulation of this method.

关键词： particle swarm optimization(PSO) BP neural network numeral recognition

作者: Xu Peng

作者单位: Faculty of Technology, Harbin University, china

会议类型: 国际会议

会议名称: 2010 International Conference on Future Information Technology(2010年未来信息技术国际会议 ICFIT 2010)

会议地点: 长沙

会议语种:英文

页码: 787-789

在线出版日期: 2010-12-14（万方平台首次上网日期，不代表论文的发表时间）