Presentation 2004/3/12
Spectral Gradient Method to Extract Character and Text Regions from General Images: Construction of A Realtime Processing System
Yoichiro BABA, Akira HIROSE,
PDF Download Page PDF download Page Link
Abstract(in Japanese) (See Japanese page)
Abstract(in English) To obtain text information from scene images, it is required to extract character/text regions included in the images before recognizing the characters. In this paper, we propose the 2-D spectrum gradient method and hybrid spectral gradient method, which is an improved version of 2-D method. We demonstrate that the system extracts the character/text regions effectively. We also show that our system can operate in real time with an improvement of the algorithm and a hardware implementation on an FPGA.
Keyword(in Japanese) (See Japanese page)
Keyword(in English) Attention / Foveation / Character-region extraction / Texture / Image indexing / Hardware
Paper # IE2003-248
Date of Issue

Conference Information
Committee IE
Conference Date 2004/3/12(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 Image Engineering (IE)
Language JPN
Title (in Japanese) (See Japanese page)
Sub Title (in Japanese) (See Japanese page)
Title (in English) Spectral Gradient Method to Extract Character and Text Regions from General Images: Construction of A Realtime Processing System
Sub Title (in English)
Keyword(1) Attention
Keyword(2) Foveation
Keyword(3) Character-region extraction
Keyword(4) Texture
Keyword(5) Image indexing
Keyword(6) Hardware
1st Author's Name Yoichiro BABA
1st Author's Affiliation Dept. of Electronic Engineering, The University of Tokyo()
2nd Author's Name Akira HIROSE
2nd Author's Affiliation Dept. of Electronic Engineering, The University of Tokyo
Date 2004/3/12
Paper # IE2003-248
Volume (vol) vol.103
Number (no) 739
Page pp.pp.-
#Pages 6
Date of Issue