Title: Learning 3D geometric features for soft-biometrics recognition [Link]

Author: Baiqiang Xia

Supervisors: Prof Mohamed Daoudi, Prof. Boulbaba Benamor

Affiliation: University of Lille 1 - Science and Technology, France

Year: 2014 

Abstract:

Soft-Biometric (gender, age, etc.) recognition has shown growing applications in different domains. Previous 2D face based studies are sensitive to illumination and pose changes, and insufficient to represent the facial morphology. To overcome these problems, this thesis employs the 3D face in Soft-Biometric recognition. Based on a Riemannian shape analysis of facial radial curves, four types of Dense Scalar Field (DSF) features are proposed, which represent the Averageness, the Symmetry, the global Spatiality and the local Gradient of 3D face. Experiments with Random Forest on the 3D FRGCv2 dataset demonstrate the effectiveness of the proposed features in Soft-Biometric recognition. Furtherly, we demonstrate the correlations of Soft-Biometrics are useful in the recognition. To the best of our knowledge, this is the first work which studies age estimation, and the correlations of Soft-Biometrics, using 3D face.