End-to-end Recovery of Human Shape and Pose

We describe Human Mesh Recovery (HMR), an end-to-end framework for reconstructing a full 3D mesh of a human body from a single RGB image. In contrast to most current methods that compute 2D or 3D joint locations, we produce a richer and more useful mesh representation that is parameterized by shape and 3D joint angles. The main objective is to minimize the reprojection loss of keypoints, which allow our model to be trained using images in-the-wild that only have ground truth 2D annotations. However, the reprojection loss alone leaves the model highly under constrained. In this work we address this problem by introducing an adversary trained to tell whether a human body parameter is real or not using a large database of 3D human meshes. We show that HMR can be trained with and without using any paired 2D-to-3D supervision. We do not rely on intermediate 2D keypoint detections and infer 3D pose and shape parameters directly from image pixels. Our model runs in real-time given a bounding box containing the person. We demonstrate our approach on various images in-the-wild and out-perform previous optimization based methods that output 3D meshes and show competitive results on tasks such as 3D joint location estimation and part segmentation.

PDF Abstract CVPR 2018 PDF CVPR 2018 Abstract

Results from the Paper


Task Dataset Model Metric Name Metric Value Global Rank Result Benchmark
3D Multi-Person Pose Estimation AGORA HMR B-NMVE 217.0 # 4
B-NMJE 226.0 # 4
B-MVE 173.6 # 4
B-MPJPE 180.5 # 4
Weakly-supervised 3D Human Pose Estimation Human3.6M Kanzawa et al. 3D Annotations No # 1
3D Human Pose Estimation Human3.6M HMR Average MPJPE (mm) 87.97 # 299
PA-MPJPE 58.1 # 103
Monocular 3D Human Pose Estimation Human3.6M HMR Use Video Sequence No # 1
Frames Needed 1 # 1
Need Ground Truth 2D Pose No # 1
3D Human Pose Estimation MPI-INF-3DHP HMR AUC 36.5 # 71
MPJPE 124.2 # 82
PA-MPJPE 89.8 # 25
PCK 72.9 # 76

Results from Other Papers


Task Dataset Model Metric Name Metric Value Rank Source Paper Compare
3D Human Pose Estimation 3DPW HMR MPJPE 130.0 # 104
Acceleration Error 37.4 # 22
3D Human Pose Estimation AGORA HMR B-NMVE 217.0 # 11
B-NMJE 226.0 # 11
B-MVE 173.6 # 11
B-MPJPE 180.5 # 11
3D Human Shape Estimation SSP-3D HMR PVE-T-SC 22.9 # 9
mIOU 69.0 # 5
3D Human Shape Estimation SSP-3D HMR(unpaired) PVE-T-SC 20.8 # 7
mIOU 61.0 # 7

Methods


No methods listed for this paper. Add relevant methods here