Presentation 2011-11-24
Fast telop character recognition without extracting character area
Satoshi SHIMADA, Jiro NAGAO, Suguru HIGASHINO,
PDF Download Page PDF download Page Link
Abstract(in Japanese) (See Japanese page)
Abstract(in English) Telop characters may be difficult to separate from the background because they are superimposed on the background image. On the other hand, character recognition methods by matching partial areas of the image to dictionary data repeatedly without separating the text and background take much processing time because of their high computational cost. This paper proposes an efficient matching method focusing on regularity of the arrangement of characters and same common features of the telop characters. First, the telop area is estimated by using density of color edge pairs. And then, midline position, the height and font type of the telop are determined by recognizing one character in the estimated telop area. Finally, each character of telop is recognized by minimum matching process. Our experiments using relatively large Gothic type telops of news TV programs confirmed that the proposed method can recognize characters at the recall and precision rate of 80 to 90 percent and at the processing time of about 0.5 second per input image.
Keyword(in Japanese) (See Japanese page)
Keyword(in English) character recognition / telop / video retrieval / edge pair
Paper # PRMU2011-109
Date of Issue

Conference Information
Committee PRMU
Conference Date 2011/11/17(1days)
Place (in Japanese) (See Japanese page)
Place (in English)
Topics (in Japanese) (See Japanese page)
Topics (in English)
Chair
Vice Chair
Secretary
Assistant

Paper Information
Registration To Pattern Recognition and Media Understanding (PRMU)
Language JPN
Title (in Japanese) (See Japanese page)
Sub Title (in Japanese) (See Japanese page)
Title (in English) Fast telop character recognition without extracting character area
Sub Title (in English)
Keyword(1) character recognition
Keyword(2) telop
Keyword(3) video retrieval
Keyword(4) edge pair
1st Author's Name Satoshi SHIMADA
1st Author's Affiliation NTT Cyber Solutions Laboratories()
2nd Author's Name Jiro NAGAO
2nd Author's Affiliation NTT Cyber Solutions Laboratories:(Present office)NTT Com
3rd Author's Name Suguru HIGASHINO
3rd Author's Affiliation NTT Cyber Solutions Laboratories
Date 2011-11-24
Paper # PRMU2011-109
Volume (vol) vol.111
Number (no) 317
Page pp.pp.-
#Pages 6
Date of Issue