Share Email Print

Proceedings Paper

2D/3D facial feature extraction
Author(s): Hatice Çinar Akakin; Albert Ali Salah; Lale Akarun; Bülent Sankur
Format Member Price Non-Member Price
PDF $17.00 $21.00

Paper Abstract

We propose and compare three different automatic landmarking methods for near-frontal faces. The face information is provided as 480x640 gray-level images in addition to the corresponding 3D scene depth information. All three methods follow a coarse-to-fine suite and use the 3D information in an assist role. The first method employs a combination of principal component analysis (PCA) and independent component analysis (ICA) features to analyze the Gabor feature set. The second method uses a subset of DCT coefficients for template-based matching. These two methods employ SVM classifiers with polynomial kernel functions. The third method uses a mixture of factor analyzers to learn Gabor filter outputs. We contrast the localization performance separately with 2D texture and 3D depth information. Although the 3D depth information per se does not perform as well as texture images in landmark localization, the 3D information has still a beneficial role in eliminating the background and the false alarms.

Paper Details

Date Published: 15 March 2006
PDF: 12 pages
Proc. SPIE 6064, Image Processing: Algorithms and Systems, Neural Networks, and Machine Learning, 60641D (15 March 2006); doi: 10.1117/12.643099
Show Author Affiliations
Hatice Çinar Akakin, Bogaziçi Univ. (Turkey)
Albert Ali Salah, Bogaziçi Univ. (Turkey)
Lale Akarun, Bogaziçi Univ. (Turkey)
Bülent Sankur, Bogaziçi Univ. (Turkey)

Published in SPIE Proceedings Vol. 6064:
Image Processing: Algorithms and Systems, Neural Networks, and Machine Learning
Nasser M. Nasrabadi; Edward R. Dougherty; Jaakko T. Astola; Syed A. Rizvi; Karen O. Egiazarian, Editor(s)

© SPIE. Terms of Use
Back to Top
Sign in to read the full article
Create a free SPIE account to get access to
premium articles and original research
Forgot your username?