首页> 外国专利> And apparatus for character pattern recognition processing method, and the recording medium and character pattern recognition processing program

And apparatus for character pattern recognition processing method, and the recording medium and character pattern recognition processing program

机译:以及用于字符模式识别处理方法的设备,记录介质和字符模式识别处理程序

摘要

PROBLEM TO BE SOLVED: To actualize high recognition performance for a character object having faintness, a stain, etc. ;SOLUTION: A deterioration decision part 1-3 count pixel connections of black to black, white to white, black to white, and white to black by scanning the whole character normalized by a preprocessing part 1-2, finds simplicity with the product of the numbers of pixel connections of black to black and white to white and complicity with the product of the numbers of pixel connections of black to white and white to black, and decides whether the character is faint or stained according to a distance value, etc., found with a deterioration decision feature standard dictionary 1-5. A deterioration correction part 1-6 uses the decision result to calculate a correction coefficient from the simplicity and complicity of a local area and corrects a feature vector using the black pixel connection length of the object character that a feature extraction part P4 extracts from a normalized character pattern by using the correction coefficient. A discrimination part 1-7 finds the distance value between the corrected feature vector and each category in a recognition feature standard dictionary 1-8 to discriminate the most similar character category as a recognition result.;COPYRIGHT: (C)2003,JPO
机译:解决的问题:为具有模糊,污点等的字符对象实现高识别性能;解决方案:劣化判定部分1-3计算黑色到黑色,白色到白色,黑色到白色和白色的像素连接通过扫描由预处理部分1-2归一化的整个字符,将其转换为黑色,发现黑色与黑色和白色至白色的像素连接数的乘积与黑色与白色的像素连接数的乘积的简单性从白色到黑色,并根据通过劣化判定特征标准字典1-5发现的距离值等来判定字符是淡淡的还是被染色的。劣化校正部1-6使用判定结果从局部的简单性和复杂性计算校正系数,并使用特征提取部P4从归一化后提取的目标字符的黑色像素连接长度来校正特征向量。通过使用校正系数的字符模式。判别部分1-7在识别特征标准字典1-8中找到校正后的特征向量与每个类别之间的距离值,以判别最相似的字符类别作为识别结果。版权所有:(C)2003,JPO

著录项

相似文献

  • 专利
  • 外文文献
  • 中文文献
获取专利

客服邮箱:kefu@zhangqiaokeyan.com

京公网安备:11010802029741号 ICP备案号:京ICP备15016152号-6 六维联合信息科技 (北京) 有限公司©版权所有
  • 客服微信

  • 服务号