基于Gabor变换的高鲁棒汉字识别新方法

被引:36
作者
王学文
丁晓青
刘长松
机构
[1] 清华大学电子工程系智能技术与系统国家重点实验室
关键词
Gabor滤波器; 字符识别;
D O I
暂无
中图分类号
TP391.43 [];
学科分类号
摘要
本文提出了针对字符图像的基于Gabor变换的汉字识别新方法。在对Gabor变换深人分析的基础上,本文针对汉字图像的统计信息,提出了一种有效的Gabor滤波器组参数优化方法;同时,对Gabor滤波器组的输出进行非线性变换,使其适应不同亮度和低质量灰度字符图像的识别。本文还改进了分块特征的抽取算法,提高了对字符细节的分辨能力 实验表明,这种特征抽取方法大大加强了识别系统抵御图像噪声、干扰。亮度变化、笔画模糊、笔画断裂以及字符形变的能力,在应用于各种低质量的二值或者灰度的印刷和脱机手写字符图像识别时,能获得较其他算法更良好的识别性能
引用
收藏
页码:1317 / 1322
页数:6
相关论文
共 7 条
[1]  
Uncertainty relation for resolution in space, spatial frequency, and orientation optimized by two-dimensional visual cortical filters. J G Daugman. Journal of the Optical Society of America A Optics Image Science and Vision . 1985
[2]  
Goal-directed evaluation of binarization methods. O D Trier,et al. IEEE Transactions on Pattern Analysis and Machine Intelligence . 1995
[3]  
A Gabor filter-based method for recognizing handwritten numbers. Y Hamamoto,et al. Pattern Recognition . 1998
[4]  
Unsupervised texture segmentation using Gabor filters. A K Jain,et al. Pattern Recognition . 1991
[5]  
Artificial neural networks for feature extraction and multi-variate data projection. J Mao,et al. IEEE ACM Transactions on Networking . 1995
[6]  
An Introduction to Digital Image Processing. W Niblack. . 1986
[7]  
Mutual information matching in multiresolution contexts. J P W Pluim,et al. Image and Vision Computing . 2001