Single-shot 3D multi-person pose estimation in complex images
作者：
Highlights：
• Multi-person 3D human pose estimation in rich and complex environments.
• 2D and 3D human joints are predicted using heatmaps and Occlusion Robust Pose Maps.
• The difficult problem of associating joints to people skeletons is managed using the associative embeddings method.
• The proposed approach results surpass single-shot methods of the state of the art on the CMU Panoptic dataset and MuPoTS-3D datasets.
• Experiments on the JTA Dataset show that complex urban scenarios with many people at different image resolution remains a challenge for our approach.
摘要
•Multi-person 3D human pose estimation in rich and complex environments.•2D and 3D human joints are predicted using heatmaps and Occlusion Robust Pose Maps.•The difficult problem of associating joints to people skeletons is managed using the associative embeddings method.•The proposed approach results surpass single-shot methods of the state of the art on the CMU Panoptic dataset and MuPoTS-3D datasets.•Experiments on the JTA Dataset show that complex urban scenarios with many people at different image resolution remains a challenge for our approach.
论文关键词：Multi-person,3D,Human pose,Deep learning
论文评审过程：Received 4 September 2019, Revised 19 June 2020, Accepted 1 July 2020, Available online 2 July 2020, Version of Record 7 January 2021.
论文官网地址：https://doi.org/10.1016/j.patcog.2020.107534
