Monocular and Stereo Methods for AAM Learning from Video

Authors: Jason Saragih and Roland Göcke

Presented by Jason Saragih at the IEEE Computer Society Conference on Computer Vision and Pattern Recognition CVPR2007, Minneapolis (MN), USA, 18-23 June 2007

Abstract

The active appearance model (AAM) is a powerful method for modeling deformable visual objects. One of the major drawbacks of the AAM is that it requires a training set of pseudo-dense correspondences over the whole database. In this work, we investigate the utility of stereo constraints for automatic model building from video. First, we propose a new method for automatic correspondence finding in monocular images which is based on an adaptive template tracking paradigm. We then extend this method to take the scene geometry into account, proposing three approaches, each accounting for the availability of the fundamental matrix and calibration parameters or the lack thereof. The performance of the monocular method was first evaluated on a pre-annotated database of a talking face. We then compared the monocular method against its three stereo extensions using a stereo database.

Download paper (458kB, PDF)
©2007 IEEE

Bibtex Entry

@INPROCEEDINGS{saragih_goecke2007,
AUTHOR = {J. Saragih and R. Goecke},
TITLE = {{Monocular and Stereo Methods for AAM Learning from Video}},
BOOKTITLE = {{Proceedings of the IEEE Computer Society Conference on Computer Vision and Pattern Recognition CVPR 2007}},
PUBLISHER = {IEEE Computer Society},
ADDRESS = {Minneapolis (MN), USA},
VOLUME = ,
PAGES = {--},
MONTH = jun,
YEAR = 2007,
NOTE = {{DOI: 10.1109/ICCV.2007.4409106}}}

[Homepage] [Research] [Publications]


(c) Roland Göcke
Last modified: Fri May 16 16:29:21 AUS Eastern Standard Time 2008