Multi-person 3D pose estimation from a single image captured by a fisheye camera

作者：

Highlights：

• We propose a novel method for multi-person 3D pose estimation from a fisheye image.

• A re-projection module is introduced to alleviate the negative impact of distortions.

• Absolute 3D poses are obtained by our method without using ground-truth camera parameters.

• We collect a new dataset taken by fisheye cameras for multi-person 3D pose estimation.

摘要

•We propose a novel method for multi-person 3D pose estimation from a fisheye image.•A re-projection module is introduced to alleviate the negative impact of distortions.•Absolute 3D poses are obtained by our method without using ground-truth camera parameters.•We collect a new dataset taken by fisheye cameras for multi-person 3D pose estimation.

论文关键词：Multi-person 3D pose estimation,Fisheye camera,Distortion,Dataset

论文评审过程：Received 16 November 2021, Revised 1 July 2022, Accepted 5 July 2022, Available online 9 July 2022, Version of Record 16 July 2022.

论文官网地址：https://doi.org/10.1016/j.cviu.2022.103505