Presentation 2003/9/20
Recovery of head orientation and position from a monocular sequence
Gregoire JOZANT, Eric VATIKIOTIS-BATESON,
PDF Download Page PDF download Page Link
Abstract(in Japanese) (See Japanese page)
Abstract(in English) This paper presents an approach for estimating 3D rigid head position and orientation from monocular, non-calibrated image sequences. The proposed method relies on a 2D/3D matching between the image-based tracking of at least two feature points on the face and a 3D rigid representation of these points. Possible 3D object motions are constrained to fit hypothetical rigid head motion trajectories relative to the torso. These modeled trajectories comprise only 3 degrees of freedom extracted by Principal Component Analysis of accurate head orientations obtained with OPTOTRAK. Thus, for the 2D tracked positions of each image, a unique estimation of head pose is computed. The estimation results have been validated using OPTOTRAK data obtained simultaneously with the video.
Keyword(in Japanese) (See Japanese page)
Keyword(in English) 3D head motion recovery / monocular video sequences / features tracking
Paper # HCS2003-15
Date of Issue

Conference Information
Committee HCS
Conference Date 2003/9/20(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 Human Communication Science (HCS)
Language JPN
Title (in Japanese) (See Japanese page)
Sub Title (in Japanese) (See Japanese page)
Title (in English) Recovery of head orientation and position from a monocular sequence
Sub Title (in English)
Keyword(1) 3D head motion recovery
Keyword(2) monocular video sequences
Keyword(3) features tracking
1st Author's Name Gregoire JOZANT
1st Author's Affiliation ATR Human Information Science Laboratories()
2nd Author's Name Eric VATIKIOTIS-BATESON
2nd Author's Affiliation University of British Columbia
Date 2003/9/20
Paper # HCS2003-15
Volume (vol) vol.103
Number (no) 328
Page pp.pp.-
#Pages 6
Date of Issue