Presentation 2012-05-18
Procrustes analysis characterized by heteroscedasticity, and its application
Yuka KOBAYASHI, Kyouko SUDO, Hiroshi KANEKO, Kazutoshi TAGAI, Mutsuo SANO,
PDF Download Page PDF download Page Link
Abstract(in Japanese) (See Japanese page)
Abstract(in English) In this paper we propose the method for matching a set of 3D data sequences using Procrustes Analysis, which is a generalized description of the linear regression model. It enables the flexible matching permitting rotation and the deformation. We also introduce a modified regression model with a term of heteroskedasticity in order to take account of the chage of variance. We applied the method to the recognition of sign language data obtained by motion capture, the result of which suggests the model with heteroskedasticity term is more robust to the change of variance due to the data property.
Keyword(in Japanese) (See Japanese page)
Keyword(in English) Procrustes analysis / Sign language / Moving image processing
Paper # IE2012-27,PRMU2012-12,MI2012-12
Date of Issue

Conference Information
Committee PRMU
Conference Date 2012/5/10(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) Procrustes analysis characterized by heteroscedasticity, and its application
Sub Title (in English)
Keyword(1) Procrustes analysis
Keyword(2) Sign language
Keyword(3) Moving image processing
1st Author's Name Yuka KOBAYASHI
1st Author's Affiliation Mitsubishi UFJ Trust Systems()
2nd Author's Name Kyouko SUDO
2nd Author's Affiliation NTT Cyber Space Laboratories
3rd Author's Name Hiroshi KANEKO
3rd Author's Affiliation Faculty of Science, TOHO University
4th Author's Name Kazutoshi TAGAI
4th Author's Affiliation NEC Nexsolutions
5th Author's Name Mutsuo SANO
5th Author's Affiliation Faculty of Science, Osaka Institute Technology University
Date 2012-05-18
Paper # IE2012-27,PRMU2012-12,MI2012-12
Volume (vol) vol.112
Number (no) 37
Page pp.pp.-
#Pages 6
Date of Issue