Presentation 2003/10/16
Adaptation of Appearance Model for Human Tracking Using Geometorical Pixel Value Distributions
Akira UTSUMI, Hirotake YAMAZOE, Nobuji TETSUTANI,
PDF Download Page PDF download Page Link
Abstract(in Japanese) (See Japanese page)
Abstract(in English) We propose a dynamic model adaptation method for appearance-based human tracking. We are investigating a vision-based system that recognizes human behavior. In such systems, various types of user information, such as human identification, motion trajectory and gestures information, are required. We use a human appearance model to extract this information. However, appearance can vary across individuals due to different clothes etc. even if they have similar poses. To reduce the effects of this problem, we perform a dynamic model adaptation method based on a statistical view-transition model. We can construct human appearance models for different poses that have proper texture information from a limited number of input images. This assists with both detection and tracking. Experimental results show the effectiveness of the proposed method.
Keyword(in Japanese) (See Japanese page)
Keyword(in English) View-Transition Model / Human Motion Recognition / Statistical Model / Object Tracking
Paper # DSP2003-110,ICD2003-108,IE2003-70
Date of Issue

Conference Information
Committee IE
Conference Date 2003/10/16(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) Adaptation of Appearance Model for Human Tracking Using Geometorical Pixel Value Distributions
Sub Title (in English)
Keyword(1) View-Transition Model
Keyword(2) Human Motion Recognition
Keyword(3) Statistical Model
Keyword(4) Object Tracking
1st Author's Name Akira UTSUMI
1st Author's Affiliation ATR Media Information Science Laboratories()
2nd Author's Name Hirotake YAMAZOE
2nd Author's Affiliation ATR Media Information Science Laboratories:Graduate School of Engineering Science, Osaka University
3rd Author's Name Nobuji TETSUTANI
3rd Author's Affiliation ATR Media Information Science Laboratories
Date 2003/10/16
Paper # DSP2003-110,ICD2003-108,IE2003-70
Volume (vol) vol.103
Number (no) 383
Page pp.pp.-
#Pages 6
Date of Issue