Title:
【発明の名称】文字列成分抽出装置
Document Type and Number:
Japanese Patent JP2733057
Kind Code:
B2
Abstract:
PURPOSE:To accurately detect character string components such as titles, authors, etc., out of documents having various formats, by storing extracting procedures adaptive to various formats and detecting and extracting the character strings based on those extracting procedures. CONSTITUTION:A control part 501 stores the extracting procedures for character string components and controls other parts based on these procedures. A contour tracking part 502 stores the minimum and maximum coordinate values of X and Y directions in a coupling component memory 506 with respect to a binary image read optically and stored in a picture memory 505 in case the contour of the coupling component of black image element is described in a broken line as a closed curve. A selection part 503 selects only the character string components stored in a certain area and stores an index of the coupling components, i.e., the result of said selection in a work memory 507. A detection part 504 having similar constitution to the part 503 outputs the one coincident with a certain condition out of the contents of the memory 506 limited by the memory 506 itself. In a comparison mode of the part 504, notice is given to one of four values of the memory 506 and the minimum or maximum value and its index of a noticed coordinate value are outputted.
Inventors:
TONO JUNICHI
FUJISAWA HIROMICHI
NAKANO YASUAKI
SHIMA YOSHIHIRO
FUJISAWA HIROMICHI
NAKANO YASUAKI
SHIMA YOSHIHIRO
Application Number:
JP20178687A
Publication Date:
March 30, 1998
Filing Date:
August 14, 1987
Export Citation:
Assignee:
KOGYO GIJUTSU INCHO
International Classes:
G06K7/10; G06K9/20; G06K9/34; (IPC1-7): G06K7/10
Domestic Patent References:
| JP62165284A | ||||
| JP6222186A |
Other References:
【文献】情報処理学会第31回 (昭和60年後期) 全国大会 2N-3 「文書理解のための画像処理方式-知的ファイリング(その3)-」