会员体验
专利管家(专利管理)
工作空间(专利管理)
风险监控(情报监控)
数据分析(专利分析)
侵权分析(诉讼无效)
联系我们
交流群
官方交流:
QQ群: 891211   
微信请扫码    >>>
现在联系顾问~
热词
    • 92. 发明专利
    • Image processor and image processing method
    • 图像处理器和图像处理方法
    • JP2005157448A
    • 2005-06-16
    • JP2003390748
    • 2003-11-20
    • Canon Incキヤノン株式会社
    • KANEDA KITAHIRO
    • G06T1/00G06K9/00G06K9/46G06T11/20G06T11/60
    • G06K9/4647G06K2209/01
    • PROBLEM TO BE SOLVED: To provide an image processor and image processing method which can improve processing efficiency when obtaining data for enabling reuse or re-edition of collection of originals such as paper documents. SOLUTION: While image data is inputted and character information included in the inputted image data is recognized sequentially, the character information included in the inputted image data are outlined by using vector data. At this time, character types of recognized character information and the outlined vector data are matched and managed. When the character information is recognized here, whether the character type of the character information concerned has been outlined in the past or not is distinguished. When distinguished as outlined in the past, the character type is controlled to be managed using the vector data at the time of being outlined in the past. COPYRIGHT: (C)2005,JPO&NCIPI
    • 要解决的问题:提供一种图像处理器和图像处理方法,其可以在获得用于使得能够重新使用或重新编辑诸如纸质文档的原件的收集的数据时提高处理效率。

      解决方案:当输入图像数据并且顺序地识别输入的图像数据中包括的字符信息时,通过使用矢量数据来概括输入的图像数据中包括的字符信息。 此时,识别的字符信息和轮廓矢量数据的字符类型被匹配和管理。 当在此识别字符信息时,区分是否已经概述了关于字符信息的字符类型。 当如以前概述的那样区分时,在过去概述时,使用向量数据来控制字符类型被管理。 版权所有(C)2005,JPO&NCIPI

    • 97. 发明专利
    • Extracting method for character feature
    • 特征提取方法
    • JPS5771079A
    • 1982-05-01
    • JP14782080
    • 1980-10-22
    • Fujitsu Ltd
    • TAKAHASHI SHIYOUJIHARADA FUMIAKIOOTA TOMOMIISOBE YUUJI
    • G06K9/46
    • G06K9/4647G06K2209/01
    • PURPOSE:To shorten the time of character recognition by reducing an feature extracting area for the character recognition by storing positions of change points simultaneously at the prescribed readout units, and reading only a unit having a change point. CONSTITUTION:The starting address of one scan is set from a CPU to a memory address counter 51 through a multiplexer 50, the number of scans is set to a scan counter 52, and the output of an adder 59 is inputted to the multiplexer 50. The storage contents of a character pattern memory 60 are read out by an address from the counter 51 to store the position of a change point in a superposition write register 54 by a readout unit. The all-0 state of the data in the memory 60 is discriminated by an all-0 discriminating circuit 61, whose result is supplied to a timing pulse generating circuit 53, which when it is judged as the change-point, supplies a set pulse to a change-point register file 55, driving a register file address counter 56. Then, only the unit having the change point is reread to reduce a feature extracting area for character recognition.
    • 目的:通过在规定的读出单元中同时存储变化点的位置,并且仅读取具有变化点的单位,通过减少用于字符识别的特征提取区域来缩短字符识别的时间。 构成:通过多路复用器50将一次扫描的起始地址从CPU设置到存储器地址计数器51,将扫描次数设置为扫描计数器52,并将加法器59的输出输入到多路复用器50。 字符图形存储器60的存储内容通过来自计数器51的地址读出,以通过读出单元将重写位置存储在叠加写入寄存器54中。 存储器60中的数据的全0状态由全0鉴别电路61鉴别,全结果被提供给定时脉冲发生电路53,当定时脉冲发生电路53被判断为改变点时,提供一个设定脉冲 变更点寄存器文件55,驱动寄存器文件地址计数器56.然后,只有具有变化点的单元被重新读取以减少用于字符识别的特征提取区域。