Title
Determining Initial Head Orientation For Real-Time Video Input
Keywords
Face detection; Face tracking; Pose estimation; Real-time video processing
Abstract
In this paper, we present a novel method for estimating face orientation suitable for use in real-time video analysis. This method uses the symmetry of generic feature points inside the region returned by a face detector instead of locating specific facial features. Finding facial features such as the eyes or the mouth generally requires computationally expensive pattern correlation operations and degrades in the presence of occlusions such as glasses or beards. Since most of the occlusions follow the face symmetry, they actually improve the results with our new method. This new method for initializing face orientation is suitable for applications such as mixed reality or input to a video game. © The University of Wolverhampton, UK.
Publication Date
1-1-2005
Publication Title
Proceedings of CGAMES 2005 - 7th International Conference on Computer Games: Artificial Intelligence, Animation, Mobile, Educational and Serious Games
Number of Pages
28-32
Document Type
Article; Proceedings Paper
Personal Identifier
scopus
Copyright Status
Unknown
Socpus ID
84906985661 (Scopus)
Source API URL
https://api.elsevier.com/content/abstract/scopus_id/84906985661
STARS Citation
Spencer, Lisa and Guha, Ratan K., "Determining Initial Head Orientation For Real-Time Video Input" (2005). Scopus Export 2000s. 4238.
https://stars.library.ucf.edu/scopus2000/4238