An Improved Manchu Character Recognition Method
S. Xu†*, G. Q. Qi‡, M. Li§, R. R. Zheng††, & C. Johnξ
†‡Information Science & Technology, Dalian Maritime University, Dalian Liaoning 116026, China,
†§††College of Information and Communication Engineering, Dalian Nationalities University,
Dalian Liaoning 116605, China
ξSchool of Engineering, University of St. Thomas, St. Paul, MN 55105-1079, USA
ABSTRACT: To improve the off-line Manchu printed character recognition rate, a method of Manchu recognition based on the letters is presented. Firstly, the preprocessing is performed to segment the Manchu letters aiming at Manchu character image. Secondly, extract the rough grid characteristics and connected domain characteristics of the Manchu letters, then using SVM and BP neural network to recognize the combination features of these ones. Finally, the grid-search method and cross-validation method are used to optimize the SVM kernel function parameters. The result of the experiment shows that the recognition rate of SVM is higher than the BP neural network, and has a better classification results.
Keywords : Off-Line Manchu character; BP neural networks; SVM; Grid-Search; Cross-Validation.
 G. Y. Zhang, Study of Offline Handwritten Manchu Character Recognition, Northeastern University, 2006.
 Z. H. Li, G. L. Gao, Extraction of Features of Mongolian Printed Character Recognition, Microcomputer, 11(2013):117-119.
 J. Z. Jia, The Research of feature selection and classifier design for Printed Offline Uygur character recognition, D. Soochow University, 2008.
 Y. L. Wang, Y. Z. Li, R. S. Wang, Application of Rough Grid in Feature Extraction of Printed Tibetan character, Sci. Technol. Eng, 18(2009): 5546-5548.
 Y. D. Song, Research of Parameter Selection for Support Vector Machine, Central China Normal University, 2013.
 Francesco Camastra, A SVM-based cursive character recognizer, Pattern Recognit, 40 (2007): 3721-3727.
 G. H. Feng,. Parameter optimizing for Support Vector Machines classification, Comput. Eng, 3, (2011):123-128.
 G. Sun, Research of Multi-class Classification Methods Based on Support Vector Machine, Dalian Maritime University, 2008.
 X. L. Wang, Z. B. Li, Identifying the Parameters of the Kernel Function in Support Vector Machines Based on the Grid-Search Method, Journal of Ocean University of Qingdao, 5(2012):859-862.