基于小波变换和DCT的字符图像特征抽取新方法

被引：3

作者：

曹建海

路长厚

机构：

[1] 山东大学机械工程学院

[2] 山东大学机械工程学院山东济南

[3] 山东济南

来源：

光电子·激光 | 2004年 / 04期

关键词：

特征抽取; 小波变换(WT); 离散余弦变换(DCT); 圆周投影; 光学字符识别;

D O I：

10.16136/j.joel.2004.04.023

中图分类号：

TP391.43 []; TP751. [];

学科分类号：

0811 ; 081101 ; 081104 ; 1405 ; 081002 ;

摘要：

从特征矢量的不变性和抗噪性角度,提出了一种基于小波变换(WT)和离散余弦变换(DCT)的字符特征抽取新方法。利用圆周投影算法,把二维的字符图像转换为一维投影。基于WT和DCT的非线性变换,克服了因变形和噪声引起的一维投影的非线性变形失真。通过对识别不同大小、方向及噪声的字符仿真实验和不同特征抽取方法的对比实验,以及对工业标牌字符的识别,表明该特征抽取方法具有尺度和旋转不变性,有较好的抗噪声能力和很好的分类性能。

引用

页码：477 / 482

页数：6

共 11 条

[1]

Texture features for DCT-coded image retrieval and classification. Huang Y L,Chang R F. IEEE Intl . Conf . On Acoustics , Speech , and Signal Processing . 1999

[2]

Image retrieval based on subband energy histograms of reordered DCT cofficients. WU Dongsheng,Wu Leuan. Signal Processing . 2002

[3]

New method of optimal smpling features for offline hardwritten Chinese character recognition. Zhang Rui,et al. International Journal of Image and Graphics . 2002

[4]

MUltispace KL for pattern representation and classification. Raffaele Cappelli,et al. IEEE Transactions on Pattern Analysis and Machine Intelligence . 2001

[5]

Recognition of a limited Chinese character set based on PCA learning jubspace algorithm. Jiang Wei-feng,Liu Ji-lin. International Journal of Image and Graphics . 2001

[6]

Fast algorithm for computing discrete cosine transform. Kok C W. IEEE Transactions on Signal Processing . 1997

[7]

Orthonormal bases of compactly supported wavelets. I. Daubechies. Communications on Pure and Applied Analysis . 1988

[8]

Orthonormal bases of compactly supported wavelets. I. Daubechies. Communications on Pure and Applied Analysis . 1988

[9]

A new normalization technique for cursive handwritten words. Alessanddro Vinciarelli,et al. Pattern Recognition . 2001

[10]

Slant estimation algorithm for OCR systems. E Kavallieratou,et al. Pattern Recognition . 2001

← 1 2 →