首页> 外文期刊>電子情報通信学会技術研究報告. 医用画像. Medical Imaging >Off-line hand-written character recognition method using on-line writing information including writing pressure features
【24h】

Off-line hand-written character recognition method using on-line writing information including writing pressure features

机译:使用包括书写压力特征的在线书写信息的离线手写字符识别方法

获取原文
获取原文并翻译 | 示例
       

摘要

In off-line handwriting character recognition, the improvement of a recognition performance to a deformed character pattern is a big subject. Although many recognition techniques are proposed, it has not been achieved sufficient recognition performance for practical use. This research examined the recognition technique strong against deformed characters, extending the conventional recognition technique by using the on-line character pattern information containing writing pressure features. In order to check the effects to off-line character recognition by using on-line character information, simple pattern matching and HMM character recognition method were extended, and the recognition system was made as an evaluation experiment. By using on-line writing features and writing pressure features, character recognition performances are increased by about 2% for simple pattern matching method, about 2% for HMM character recognition as a result of the recognition experiment using the common handwriting English character database.
机译:在离线手写字符识别中,对变形的字符图案的识别性能的改善是很大的课题。尽管提出了许多识别技术,但尚未获得足够的识别性能以供实际使用。这项研究研究了对变形字符具有强力的识别技术,通过使用包含书写压力特征的在线字符模式信息扩展了传统的识别技术。为了验证利用在线字符信息对离线字符识别的影响,扩展了简单模式匹配和HMM字符识别方法,并建立了识别系统作为评估实验。通过使用在线书写功能和书写压力功能,通过使用普通手写英文字符数据库进行的识别实验,对于简单的模式匹配方法,字符识别性能提高了约2%,对于HMM字符识别,性能提高了约2%。

著录项

相似文献

  • 外文文献
  • 中文文献
  • 专利
获取原文

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