Head Pose Estimation from Passive Stereo Images

Michael D. Breitenstein, Jeppe Jensen, Carsten Høilund, Thomas B. Moeslund, Luc Van Gool

Research output: Contribution to journalConference article in JournalResearchpeer-review

11 Citations (Scopus)

Abstract

We present an algorithm to estimate the 3D pose (location and orientation) of a previously unseen face from low-quality range images. The algorithm generates many pose candidates from a signature to find the nose tip based on local shape, and then evaluates each candidate by computing an error function. Our algorithm incorporates 2D and 3D cues to make the system robust to low-quality range images acquired by passive stereo systems. It handles large pose variations (of ±90 ° yaw and ±45 ° pitch rotation) and facial variations due to expressions or accessories. For a maximally allowed error of 30°, the system achieves an accuracy of 83.6%.
Original languageEnglish
Book seriesLecture Notes in Computer Science
Volume5575
Pages (from-to)219-228
Number of pages9
ISSN0302-9743
DOIs
Publication statusPublished - 2009
EventScandinavian Conference on Image Analysis - Oslo, Norway
Duration: 15 Jun 200918 Jun 2009
Conference number: LNCS 5575

Conference

ConferenceScandinavian Conference on Image Analysis
NumberLNCS 5575
CountryNorway
CityOslo
Period15/06/200918/06/2009

Fingerprint

Pose Estimation
Range Image
Error function
Accessories
Signature
Face
Computing
Evaluate
Estimate

Cite this

Breitenstein, Michael D. ; Jensen, Jeppe ; Høilund, Carsten ; Moeslund, Thomas B. ; Gool, Luc Van. / Head Pose Estimation from Passive Stereo Images. In: Lecture Notes in Computer Science. 2009 ; Vol. 5575. pp. 219-228.
@inproceedings{2eac5fb0b73211debd73000ea68e967b,
title = "Head Pose Estimation from Passive Stereo Images",
abstract = "We present an algorithm to estimate the 3D pose (location and orientation) of a previously unseen face from low-quality range images. The algorithm generates many pose candidates from a signature to find the nose tip based on local shape, and then evaluates each candidate by computing an error function. Our algorithm incorporates 2D and 3D cues to make the system robust to low-quality range images acquired by passive stereo systems. It handles large pose variations (of ±90 ° yaw and ±45 ° pitch rotation) and facial variations due to expressions or accessories. For a maximally allowed error of 30°, the system achieves an accuracy of 83.6{\%}.",
author = "Breitenstein, {Michael D.} and Jeppe Jensen and Carsten H{\o}ilund and Moeslund, {Thomas B.} and Gool, {Luc Van}",
note = "Titel: Proceedings of the 16th Scandinavian Conference on Image Analysis. Oversat titel: Oversat undertitel: Forlag: Springer ISBN (Trykt): 9783642022296 ISBN (Elektronisk): Publikationsserier: Lecture Notes In Computer Science, Springer, 0302-9743 Image Processing, Computer Vision, Pattern Recognition, and Graphics, Springer, 5575",
year = "2009",
doi = "10.1007/978-3-642-02230-2_23",
language = "English",
volume = "5575",
pages = "219--228",
journal = "Lecture Notes in Computer Science",
issn = "0302-9743",
publisher = "Physica-Verlag",

}

Head Pose Estimation from Passive Stereo Images. / Breitenstein, Michael D.; Jensen, Jeppe; Høilund, Carsten; Moeslund, Thomas B.; Gool, Luc Van.

In: Lecture Notes in Computer Science, Vol. 5575, 2009, p. 219-228.

Research output: Contribution to journalConference article in JournalResearchpeer-review

TY - GEN

T1 - Head Pose Estimation from Passive Stereo Images

AU - Breitenstein, Michael D.

AU - Jensen, Jeppe

AU - Høilund, Carsten

AU - Moeslund, Thomas B.

AU - Gool, Luc Van

N1 - Titel: Proceedings of the 16th Scandinavian Conference on Image Analysis. Oversat titel: Oversat undertitel: Forlag: Springer ISBN (Trykt): 9783642022296 ISBN (Elektronisk): Publikationsserier: Lecture Notes In Computer Science, Springer, 0302-9743 Image Processing, Computer Vision, Pattern Recognition, and Graphics, Springer, 5575

PY - 2009

Y1 - 2009

N2 - We present an algorithm to estimate the 3D pose (location and orientation) of a previously unseen face from low-quality range images. The algorithm generates many pose candidates from a signature to find the nose tip based on local shape, and then evaluates each candidate by computing an error function. Our algorithm incorporates 2D and 3D cues to make the system robust to low-quality range images acquired by passive stereo systems. It handles large pose variations (of ±90 ° yaw and ±45 ° pitch rotation) and facial variations due to expressions or accessories. For a maximally allowed error of 30°, the system achieves an accuracy of 83.6%.

AB - We present an algorithm to estimate the 3D pose (location and orientation) of a previously unseen face from low-quality range images. The algorithm generates many pose candidates from a signature to find the nose tip based on local shape, and then evaluates each candidate by computing an error function. Our algorithm incorporates 2D and 3D cues to make the system robust to low-quality range images acquired by passive stereo systems. It handles large pose variations (of ±90 ° yaw and ±45 ° pitch rotation) and facial variations due to expressions or accessories. For a maximally allowed error of 30°, the system achieves an accuracy of 83.6%.

U2 - 10.1007/978-3-642-02230-2_23

DO - 10.1007/978-3-642-02230-2_23

M3 - Conference article in Journal

VL - 5575

SP - 219

EP - 228

JO - Lecture Notes in Computer Science

JF - Lecture Notes in Computer Science

SN - 0302-9743

ER -