Camera Distance-aware Top-down Approach for 3D Multi-person Pose Estimation from a Single RGB Image

ICCV 2019  ยท  Gyeongsik Moon, Ju Yong Chang, Kyoung Mu Lee ยท

Although significant improvement has been achieved recently in 3D human pose estimation, most of the previous methods only treat a single-person case. In this work, we firstly propose a fully learning-based, camera distance-aware top-down approach for 3D multi-person pose estimation from a single RGB image. The pipeline of the proposed system consists of human detection, absolute 3D human root localization, and root-relative 3D single-person pose estimation modules. Our system achieves comparable results with the state-of-the-art 3D single-person pose estimation models without any groundtruth information and significantly outperforms previous 3D multi-person pose estimation methods on publicly available datasets. The code is available in https://github.com/mks0601/3DMPPE_ROOTNET_RELEASE , https://github.com/mks0601/3DMPPE_POSENET_RELEASE.

PDF Abstract ICCV 2019 PDF ICCV 2019 Abstract

Results from the Paper


 Ranked #1 on Monocular 3D Human Pose Estimation on Human3.6M (Use Video Sequence metric)

     Get a GitHub badge
Task Dataset Model Metric Name Metric Value Global Rank Uses Extra
Training Data
Result Benchmark
3D Human Pose Estimation 3D Poses in the Wild Challenge RootNet MPJPE 84.28 # 4
MPJAE 21.25 # 2
Root Joint Localization Human3.6M 3DMPPE_ROOTNET MRPE 120.0 # 2
3D Absolute Human Pose Estimation Human3.6M RootNet MRPE 120.0 # 2
3D Human Pose Estimation Human3.6M PoseNet (GTi) Average MPJPE (mm) 53.3 # 205
Monocular 3D Human Pose Estimation Human3.6M Moon et. al. Use Video Sequence No # 1
Frames Needed 1 # 1
Need Ground Truth 2D Pose No # 1
3D Human Pose Estimation Human3.6M PoseNet Average MPJPE (mm) 54.4 # 212
3D Multi-Person Pose Estimation (root-relative) MuPoTS-3D 3DMPPE_POSENET 3DPCK 81.8 # 14
3D Multi-Person Pose Estimation (absolute) MuPoTS-3D 3DMPPE_POSENET 3DPCK 31.5 # 13

Methods


No methods listed for this paper. Add relevant methods here