A constrained hybrid optimization algorithm for morphable appearance models

Authors:
Cuiping Zhang;Fernand S. Cohen
Affiliations:
Electrical and Computer Engineering department, Drexel University, Philadelphia, PA;Electrical and Computer Engineering department, Drexel University, Philadelphia, PA
Venue:
EMMCVPR'05 Proceedings of the 5th international conference on Energy Minimization Methods in Computer Vision and Pattern Recognition
Year:
2005

Citing 8
Cited 0

Active shape models—their training and application

Computer Vision and Image Understanding
Active Appearance Models

IEEE Transactions on Pattern Analysis and Machine Intelligence
The FERET Evaluation Methodology for Face-Recognition Algorithms

CVPR '97 Proceedings of the 1997 Conference on Computer Vision and Pattern Recognition (CVPR '97)
Coding Facial Expressions with Gabor Wavelets

FG '98 Proceedings of the 3rd. International Conference on Face & Gesture Recognition
Face Shape Extraction and Recognition Using 3D Morphing and Distance Mapping

FG '00 Proceedings of the Fourth IEEE International Conference on Automatic Face and Gesture Recognition 2000
Multidimensional Morphable Models

ICCV '98 Proceedings of the Sixth International Conference on Computer Vision
Active Appearance Models Revisited

International Journal of Computer Vision
Active morphable model: an efficient method for face analysis

FGR' 04 Proceedings of the Sixth IEEE international conference on Automatic face and gesture recognition

Quantified Score

Hi-index	0.00

Visualization

Abstract

In this paper, we propose a constrained hybrid optimization algorithm that incorporates several shape constraints into a gradient descent procedure using a novel unbiased cost function. Shape constraints are heuristically derived from face images where the face shape can be directly estimated based on ”motion” analysis. To better locate face contour points regardless of the background, local projection models are used. Experiments show that our algorithm benefits significantly from these shape constraints and achieves a much higher convergent rate compared to the inverse compositional optimization algorithm. We test our algorithm on different face databases, and demonstrate its robustness in presence of various illuminations, background patterns, as well as variations in face expressions.