Accurate Head Pose Estimation Based on Multi-Stage Regression

Yinchuan Liu, Yufei Gong, Zheng Lu, Xuetao ZhangICIP 2022 · OpenReview

Abstract

This paper proposes a method for head pose estimation from a single image. We employ a multi-stage regression strategy. To overcome the discontinuity of Euler angles and quaternions and avoid the additional constraints required to directly regress the rotation matrix, we apply a continuous 6D representation to the head pose estimation problem. Each stage of the network regresses two 1 × 3 vectors, which are then transformed into a 3 × 3 rotation matrix by this continuous 6D representation. To better perceive the difference in rotation angles, we adopt the Riemann distance to measure the closeness between the network-estimated rotation matrix and the ground truth rotation matrix corresponding to the head pose. Experiments show that our method achieves the state-of-the-art on BIWI dataset and performs favorably on AFLW2000 dataset.
BibTeX
@inproceedings{liu2022accurate,
  author = {Yinchuan Liu and Yufei Gong and Zheng Lu and Xuetao Zhang},
  title = {{Accurate Head Pose Estimation Based on Multi-Stage Regression}},
  booktitle = {ICIP},
  year = {2022},
  url = {https://openreview.net/forum?id=PzsG1LvXCN},
}

Loading related content...

Figure