| 基于有效鉴别特征抽取的有限汉字集识别 |
| |
| 引用本文： | 马俊莉,莫玉龙,王明祥.基于有效鉴别特征抽取的有限汉字集识别[J].上海大学学报(自然科学版),2002,8(1):28-30,34. |
| |
| 作者姓名： | 马俊莉 莫玉龙 王明祥 |
| |
| 作者单位： | 上海大学,通信与信息工程学院,上海,200072 |
| |
| 基金项目： | 国家自然科学基金资助项目 ( 6 9875 0 0 9) |
| |
| 摘 要： | 在模式识别领域，基于Fisher鉴别准则函数的最佳鉴别矢量集技术有着重大的影响，根据统计不相关性最优鉴别变换，提出了在汽车牌照二值化图像中抽取汉字字符的有效鉴别特征方法，并将其应用于沪宁高速公路收费口处实拍的车牌图像库中，其识别率达97.13%，实验结果表明，此文所提出的牌照图中汉字字符集的有效鉴别特征方法是有效的。
|
| 关 键 词： | 特征抽取 最佳鉴别向量 汉字识别 图像处理 模式识别 Fisher鉴别准则函数 有效鉴别特征方法 |
| 文章编号： | 1007-2861(2002)01-0028-03 |
Recognition of Limited Chinese Character Set Based on Optimal Discriminant Features Extraction |
| |
| MA Jun-li,MO Yu-long,WANG Ming-xiang.Recognition of Limited Chinese Character Set Based on Optimal Discriminant Features Extraction[J].Journal of Shanghai University(Natural Science),2002,8(1):28-30,34. |
| |
| Authors: | MA Jun-li MO Yu-long WANG Ming-xiang |
| |
| Abstract: | Optimal discriminant vectors based on Fisher's discriminant criterion have significant influence on the area of pattern recognition. In this paper, we propose a method to extract uncorrelated optimal discriminant features of Chinese characters in binarized automobile license plate image according to the uncorrelated optimal discriminant transform. The method is applied to recognize Chinese characters in automobile license plate database of Shanghai-Nanjing highway. The recognition rate is 97.13%. Experimental results show that the uncorrelated optimal discriminant features of Chinese characters in automobile license plate image is effective. |
| |
| Keywords: | feature extraction optimal discriminant vectors Chinese characters recognition Image processing |
| 本文献已被 CNKI 维普 万方数据 等数据库收录！ |
