loading page

EgoFish3D: Egocentric 3D Pose Estimation from a Fisheye Camera via Self-Supervised Learning
  • +3
  • Yuxuan Liu ,
  • Jianxin Yang ,
  • Xiao Gu ,
  • Yijun Chen ,
  • Yao Guo ,
  • Guang-Zhong Yang
Yuxuan Liu
Institute of Medical Robotics, Institute of Medical Robotics

Corresponding Author:[email protected]

Author Profile
Jianxin Yang
Author Profile
Yijun Chen
Author Profile
Guang-Zhong Yang
Author Profile


Egocentric vision has a wide range of applications for human-centric activity recognition. However, the use of the egocentric fisheye camera allows wide angle coverage but image distortion is introduced along with strong human body self-occlusion, which can impose significant challenges in data processing and model reconstruction. Unlike previous work only leveraging synthetic data for model training, this paper first presents a new real-world EgoCentric Human Action (ECHA) dataset. By using the self-supervised learning under multi-view constraints, we propose a simple yet effective framework, namely EgoFish3D, for egocentric 3D pose estimation from a single image in different real-world scenarios.
2023Published in IEEE Transactions on Multimedia on pages 1-12. 10.1109/TMM.2023.3242551