Head Pose Estimation using Fisher Manifold Learning

  • Authors:
  • Longbin Chen;Lei Zhang;Yuxiao Hu;Mingjing Li;Hongjiang Zhang

  • Affiliations:
  • -;-;-;-;-

  • Venue:
  • AMFG '03 Proceedings of the IEEE International Workshop on Analysis and Modeling of Faces and Gestures
  • Year:
  • 2003

Quantified Score

Hi-index 0.00

Visualization

Abstract

In this paper, we proposed a new learning strategy for head poseestimation. Our approach uses nonlinear interpolation to estimatethe head pose using the learning result from face images of twohead poses. Advantage of our method to regression method is that itonly requires training images of two head poses and bettergeneralization ability. It outperforms existed methods, such asregression and multi-classclassification method, on both synthesisand real face images. Average head pose estimation error of yawrotation is about 4°, which proves that our method is effectivein head pose estimation.