Login| Sign Up| Help| Contact|

Patent Searching and Data


Title:
テキスト検出の方法及び装置
Document Type and Number:
Japanese Patent JP5796107
Kind Code:
B2
Abstract:
PROBLEM TO BE SOLVED: To provide a method and device for text detection in a natural image and a method and system for text information extraction.SOLUTION: A method for detecting a text region in an image including at least one connected component (CC) includes: a TBR detection step 310 for detecting at least one text background region (TBR) from an image; a CC filtering step 320 for ensuring at least one candidate text CC by filtering at least one CC; and a CC grouping step 330 for forming at least one CC group by grouping at least one candidate text CC on the basis of the TBR detected in the TBR detection step and generating at least one text region on the basis of at least one CC group. A location of a text in an image can be promptly and/or accurately shown, and as a result, performance in text detection increases.

Inventors:
Mar Wenfor
Tsao Hai Luo
Application Number:
JP2014103652A
Publication Date:
October 21, 2015
Filing Date:
May 19, 2014
Export Citation:
Click for automatic bibliography generation   Help
Assignee:
Canon Inc
International Classes:
G06T7/40
Domestic Patent References:
JP11238098A
JP7093476A
JP2009123206A
JP2004127203A
JP11239363A
JP2010067223A
JP2010277227A
JP2011139366A
JP2012222581A
Foreign References:
US20110069885
Other References:
北田 英樹、若原 徹,“エッジ成分の空間的配置に着目した情景画像からの文字列抽出”,情報処理学会 シンポジウム 画像の認識・理解シンポジウム(MIRU) 2011,日本,情報処理学会,2011年 7月20日,pp.1502-1507
Attorney, Agent or Firm:
Yasunori Otsuka
Shiro Takayanagi
Yasuhiro Otsuka
Shuji Kimura
Osamu Shimoyama
Nagakawa Yukimitsu



 
Previous Patent: キャリーケース

Next Patent: CRUSHER