Learning confidence transformation for handwritten Chinese text recognition

CORC > 厦门大学 > 数学科学－已发表论文

	Learning confidence transformation for handwritten Chinese text recognition
	Wang, Da-Han ; Liu, Cheng-Lin ; Wang DH(王大寒)
刊名	http://dx.doi.org/10.1007/s10032-013-0214-3
	2014
关键词	Learning systems
英文摘要	Handwritten text recognition systems commonly combine character classification confidence scores and context models for evaluating candidate segmentation-recognition paths, and the classification confidence is usually optimized at character level. In this paper, we investigate into different confidence-learning methods for handwritten Chinese text recognition and propose a string-level confidence-learning method, which estimates confidence parameters by directly optimizing the performance of character string recognition. We first compare the performances of parametric (class-dependent and class-independent parameters) and nonparametric (isotonic regression) confidence-learning methods. Then, we propose two regularized confidence estimation methods and particularly, a string-level confidence-learning method under the minimum classification error criterion. In experiments of online handwritten Chinese text recognition, the string-level confidence-learning method is shown to effectively improve the string recognition performance. Using three character classifiers, the character correct rates are improved from 92.39, 90.24 and 88.69 % to 92.76, 90.91 and 89.93 %, respectively. ? 2013 Springer-Verlag Berlin Heidelberg.
语种	英语
出版者	Springer Verlag
内容类型	期刊论文
源URL	[http://dspace.xmu.edu.cn/handle/2288/91419]
专题	数学科学－已发表论文
推荐引用方式 GB/T 7714	Wang, Da-Han,Liu, Cheng-Lin,Wang DH. Learning confidence transformation for handwritten Chinese text recognition[J]. http://dx.doi.org/10.1007/s10032-013-0214-3,2014.
APA	Wang, Da-Han,Liu, Cheng-Lin,&王大寒.(2014).Learning confidence transformation for handwritten Chinese text recognition.http://dx.doi.org/10.1007/s10032-013-0214-3.
MLA	Wang, Da-Han,et al."Learning confidence transformation for handwritten Chinese text recognition".http://dx.doi.org/10.1007/s10032-013-0214-3 (2014).