3D Aided Face Recognition across Pose Variations - Archive ouverte HAL Access content directly
Book Sections Year : 2012

3D Aided Face Recognition across Pose Variations

Di Huang
  • Function : Author
Yunhong Wang
  • Function : Author
Liming Chen

Abstract

Recently, 3D aided face recognition, concentrating on improving performance of 2D techniques via 3D data, has received increasing attention due to its wide application potential in real condition. In this paper, we present a novel 3D aided face recognition method that can deal with the probe images in different viewpoints. It first estimates the face pose based on the Random Regression Forest, and then rotates the 3D face models in the gallery set to that of the probe pose to generate specific gallery sample for matching, which largely reduces the influence of head pose variations. Experiments are carried out on a subset of the FRGC v1.0 database, and the achieved performance clearly highlights the effectiveness of the proposed method.
No file

Dates and versions

hal-01353184 , version 1 (10-08-2016)

Identifiers

Cite

Wuming Zhang, Di Huang, Yunhong Wang, Liming Chen. 3D Aided Face Recognition across Pose Variations. Biometric Recognition, Springer, pp.58-66, 2012, ⟨10.1007/978-3-642-35136-5_8⟩. ⟨hal-01353184⟩
61 View
0 Download

Altmetric

Share

Gmail Facebook X LinkedIn More