发明名称 AUTOMATIC EXTRACTING PROCESSING SYSTEM OF TEXT STRUCTURE
摘要 PURPOSE:To record automatically a document mixed with character and diagram in an optimum form by discriminating automatically characters and a diagram according to the relation of connection in longitudinal and lateral directions in the document, applying character recognition to a character region and inputting a code. CONSTITUTION:Fourier transformation is applied in longitudinal and lateral directions in an input device taking upper left part of a document as an origin. Then, a pattern having regularity is separated as a character region and a part without regularity is separated as a diagram region. Then, the size of character string and shape are classified depending on the characteristic as to the character region, further each character is applied for character recognition thereby attaining coding. Moreover, codes on a diagram region are converted into bit image and stored. Then, information such as size and location of the character region are added and recorded on a recording medium.
申请公布号 JPS59168762(A) 申请公布日期 1984.09.22
申请号 JP19830041902 申请日期 1983.03.14
申请人 NIPPON DENSHIN DENWA KOSHA 发明人 NAKAMURA YOSHIAKI;SUZUKI HAJIME;SAKAI HIROSHI
分类号 G06K9/34;G06K9/20;G06K9/36;G06T7/00;H04N1/00;H04N1/40 主分类号 G06K9/34
代理机构 代理人
主权项
地址