• 主题
高级搜索  >
历史搜索 清空历史
首页> 外文会议>SIREN;Italian workshop on neural nets;Annual meeting of the Italian Neural Network Society >Writer Recognition by Means of Stroke Categorization based on Self-Organizing Maps
【24h】

Writer Recognition by Means of Stroke Categorization based on Self-Organizing Maps

机译：基于自组织映射的笔划分类识别作家

获取原文
页面导航
  • 摘要
  • 著录项
  • 相似文献
  • 相关主题

摘要

In this paper a biometric recognition system based on handwritten words is presented. The system relies on a pair of catalogues of strokes that are built unsupervisedly by means of self-organizing maps. One of the catalogues categorizes pen-down strokes (the strokes executed exerting pressure on the writing surface) while the other categorizes pen-up strokes (in-air movements of the hand performed while transitioning from one pen-down stroke to the next). These catalogues allow mapping sequences of strokes into sequences of integers. The latter, much simpler sequences, can be effectively compared by means of dynamic time warping, taking advantage of the neighboring properties exhibited by self-organizing maps. The system yields considerably good results both in identification (95.6% accuracy) and in verification (1.57 % error) when tested with 320 users and one handwritten word.
机译：本文提出了一种基于手写单词的生物识别系统。该系统依赖于通过自组织图无监督地构建的一对笔画目录。其中一个目录将笔向下笔划归类（执行在笔表面上施加压力的笔划），而另一个目录则将笔向上笔划归类（在从一个笔向下笔划过渡到下一个笔划时所执行的手的空中运动）。这些目录允许将笔划序列映射为整数序列。后者，更简单的序列，可以利用动态时间扭曲有效地进行比较，并利用自组织图展示的相邻属性。当使用320个用户和一个手写单词进行测试时，该系统在识别（准确度为95.6％）和验证（错误为1.57％）方面均产生相当不错的结果。

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