TL;DR
This paper introduces an adversarially trained prior for SMPL parameters that ensures physically plausible human poses, improving 3D reconstruction and pose estimation from images over existing methods.
Contribution
We propose a novel adversarial prior for SMPL parameters that better constrains pose estimates, outperforming VAE-based approaches in 3D human reconstruction tasks.
Findings
Adversarial prior covers real-data pose diversity
Improves 3D reconstruction from 2D keypoints
Outperforms VAE-based constraints in pose estimation
Abstract
The Skinned Multi-Person Linear (SMPL) model can represent a human body by mapping pose and shape parameters to body meshes. This has been shown to facilitate inferring 3D human pose and shape from images via different learning models. However, not all pose and shape parameter values yield physically-plausible or even realistic body meshes. In other words, SMPL is under-constrained and may thus lead to invalid results when used to reconstruct humans from images, either by directly optimizing its parameters, or by learning a mapping from the image to these parameters. In this paper, we therefore learn a prior that restricts the SMPL parameters to values that produce realistic poses via adversarial training. We show that our learned prior covers the diversity of the real-data distribution, facilitates optimization for 3D reconstruction from 2D keypoints, and yields better pose estimates…
Peer Reviews
No public reviews on file for this paper yet. If you reviewed it on a platform where reviews are public (OpenReview, ICLR, NeurIPS, ICML), you can paste yours below so the community can read it here.
Code & Models
Videos
No videos yet. Explain this paper in a talk, walkthrough, or lecture? Add one.
