Presentation 2008-03-13
Handwritten Character Distinction Inspired by Human Vision System
Junpei KOYAMA, Masahiro KATO, Akira HIROSE,
PDF Download Page PDF download Page Link
Abstract(in Japanese) (See Japanese page)
Abstract(in English) We deal with distinction between handwritten and machine-printed characters in document images. Current distinction techniques need extraction of characters and text lines from document images. However, the extraction is difficult if document images have handwritten characters. To solve the problem, we propose a distinction method inspired by human vision system. We transform every local region in document images into frequency domain data, and extract feature values which include fluctuations caused by handwriting. For the distinction, we employ a multilayer perceptron which learns the feature values. Experimental results show that our method doesn't need to locate characters or text lines. In addition, we analyze condition for the distinction experimentally.
Keyword(in Japanese) (See Japanese page)
Keyword(in English) Distinction between handwritten and machine-printed characters / Human vision system / Two dimensional Fourier transform / Power spectrum / Multilayer perceptron
Paper # NC2007-151
Date of Issue

Conference Information
Committee NC
Conference Date 2008/3/5(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 Neurocomputing (NC)
Language JPN
Title (in Japanese) (See Japanese page)
Sub Title (in Japanese) (See Japanese page)
Title (in English) Handwritten Character Distinction Inspired by Human Vision System
Sub Title (in English)
Keyword(1) Distinction between handwritten and machine-printed characters
Keyword(2) Human vision system
Keyword(3) Two dimensional Fourier transform
Keyword(4) Power spectrum
Keyword(5) Multilayer perceptron
1st Author's Name Junpei KOYAMA
1st Author's Affiliation The University of Tokyo()
2nd Author's Name Masahiro KATO
2nd Author's Affiliation Fuji Xerox Co., Ltd.
3rd Author's Name Akira HIROSE
3rd Author's Affiliation The University of Tokyo
Date 2008-03-13
Paper # NC2007-151
Volume (vol) vol.107
Number (no) 542
Page pp.pp.-
#Pages 6
Date of Issue