首页> 外国专利> FLAW LEARNING APPARATUS, FLAW LEARNING METHOD, AND COMPUTER PROGRAM

FLAW LEARNING APPARATUS, FLAW LEARNING METHOD, AND COMPUTER PROGRAM

机译:辅助学习装置,辅助学习方法和计算机程序

摘要

PROBLEM TO BE SOLVED: To possibly and accurately inspect flaws without regularly collecting teacher data.;SOLUTION: If learning models A-C are updated, it automatically determines to implement which of learning with a teacher for applying "correct answer data regarding flaw types" configured by an inspector to flaw data and generating the learning model, or learning without the teacher for applying "data regarding flaw types" obtained by applying the learning model to the flaw data and generating the learning model based on the number of partial feature value spaces having no change before and after an update among the partial feature value spaces included in the learning models A-C.;COPYRIGHT: (C)2010,JPO&INPIT
机译:解决的问题:在不定期收集教师数据的情况下可能且准确地检查缺陷。;解决方案:如果更新了学习模型AC,它将自动确定与教师一起实施哪种学习,以应用由配置的“关于缺陷类型的正确答案数据”检查员检查缺陷数据并生成学习模型,或者在没有老师的情况下进行学习,以通过将学习模型应用于缺陷数据并基于不具有特征的部分特征值空间的数量生成学习模型而获得“关于缺陷类型的数据”学习模型AC中包含的部分特征值空间在更新前后的变化。版权所有:(C)2010,JPO&INPIT

著录项

  • 公开/公告号JP2009265082A

    专利类型

  • 公开/公告日2009-11-12

    原文格式PDF

  • 申请/专利权人 NIPPON STEEL CORP;

    申请/专利号JP20090005124

  • 发明设计人 KANEKO SHUNICHI;KOBAYASHI TAKAMICHI;

    申请日2009-01-13

  • 分类号G01N21/88;G01N21/892;G06T7/00;

  • 国家 JP

  • 入库时间 2022-08-21 19:05:50

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