Presentation 2010-08-05
Learning Using Synthetic Hand Images for Fingerspelling Recognition
Kentaro TAJI, Yasuhiro OHKAWA, Nobuko KATO, Kazuhiro FUKUI,
PDF Download Page PDF download Page Link
Abstract(in Japanese) (See Japanese page)
Abstract(in English) The communication with fingerspellings is an important mean for hearing impaired people. It consists of 46 kinds of hand-shapes. In this paper, we consider non-motional fingerspellings of them as targets to be classified. We built the recognition system of these fingerspellings based on statistical pattern recognition method, Mutual Subspace Method (MSM). A lot of learning samples are required to derive the effectiveness of the MSM. However, the collection of such various images of finger spellings is difficult. To overcome this problem, we propose the framework of learning system by using hand images synthesized by Computer Graphics techniques. The experiments show that effectiveness of synthetic hand images is the same or more than that of real hand images.
Keyword(in Japanese) (See Japanese page)
Keyword(in English) Fingerspelling / Pattern Recognition / Mutual Subspace Method / Learning Sample / Computer Graphics
Paper # WIT2010-18
Date of Issue

Conference Information
Committee WIT
Conference Date 2010/7/29(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 Well-being Information Technology(WIT)
Language JPN
Title (in Japanese) (See Japanese page)
Sub Title (in Japanese) (See Japanese page)
Title (in English) Learning Using Synthetic Hand Images for Fingerspelling Recognition
Sub Title (in English)
Keyword(1) Fingerspelling
Keyword(2) Pattern Recognition
Keyword(3) Mutual Subspace Method
Keyword(4) Learning Sample
Keyword(5) Computer Graphics
1st Author's Name Kentaro TAJI
1st Author's Affiliation Graduate School of Systems and Information Engineering, University of Tsukuba()
2nd Author's Name Yasuhiro OHKAWA
2nd Author's Affiliation Graduate School of Systems and Information Engineering, University of Tsukuba
3rd Author's Name Nobuko KATO
3rd Author's Affiliation National University Corporation Tsukuba University of Technology
4th Author's Name Kazuhiro FUKUI
4th Author's Affiliation Graduate School of Systems and Information Engineering, University of Tsukuba
Date 2010-08-05
Paper # WIT2010-18
Volume (vol) vol.110
Number (no) 164
Page pp.pp.-
#Pages 6
Date of Issue