Supervised Principal Geodesic Analysis on Facial Surface Normals for Gender Classification

Authors:
Jing Wu;William A. Smith;Edwin R. Hancock
Affiliations:
Department of Computer Science, The University of York, York, UK YO10 5DD;Department of Computer Science, The University of York, York, UK YO10 5DD;Department of Computer Science, The University of York, York, UK YO10 5DD
Venue:
SSPR & SPR '08 Proceedings of the 2008 Joint IAPR International Workshop on Structural, Syntactic, and Statistical Pattern Recognition
Year:
2008

Citing 6
Cited 1

A morphable model for the synthesis of 3D faces

Proceedings of the 26th annual conference on Computer graphics and interactive techniques
Learning Gender with Support Faces

IEEE Transactions on Pattern Analysis and Machine Intelligence
A new framework for grayscale and colour non-lambertian shape-from-shading

ACCV'07 Proceedings of the 8th Asian conference on Computer vision - Volume Part II
Weighted principal geodesic analysis for facial gender classification

CIARP'07 Proceedings of the Congress on pattern recognition 12th Iberoamerican conference on Progress in pattern recognition, image analysis and applications
Multimodal facial gender and ethnicity identification

ICB'06 Proceedings of the 2006 international conference on Advances in Biometrics
Robust linear dimensionality reduction

IEEE Transactions on Visualization and Computer Graphics

Gender discriminating models from facial surface normals

Pattern Recognition

Quantified Score

Hi-index	0.00

Visualization

Abstract

In this paper, we perform gender classification based on facial surface normals (facial needle-maps). We improve our previous work in [6] by using a non-Lambertian Shape-from-Shading (SFS) method to recover the surface normals, and develop a novel supervised principal geodesic analysis (PGA) to parameterize the facial needle-maps. Experimental results demonstrate the feasibility of gender classification based on facial needle-maps, and shows that incorporating pairwise relationships between the labeled data improves the gender discriminating powers in the leading PGA eigenvectors and gender classification accuracy.