RECORD
Adversarial Parametric Pose Prior
Computer Vision and Pattern Recognition · Submitted 8 Dec 2021
abstract compiled · 2026-09-23 · no language-model review
The authors' main claim 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 when used for regression from images.
The compiled record
The paper's own sentences, arranged. Each label is drawn from a closed vocabulary; the number beside it is the probability the classifier gave that label. Numbers in the tables are the authors', checked by code. No sentence in this record was written by a model.
Open the compiled record
The abstract, sentence by sentence
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 when used for regression from images. We found that the prior based on spherical distribution gets the best results. Furthermore, in all these tasks, it outperforms the state-of-the-art VAE-based approach to constraining the SMPL parameters.
background contribution method result interpretation main claim
| # | sentence | role | the authors' own work | states the main conclusion | strength |
|---|---|---|---|---|---|
| s0 | The Skinned Multi-Person Linear (SMPL) model can represent a human body by mapping pose and shape parameters to body meshes. | background 1.00 | 0.05 | 0.02 | asserted 1.00 |
| s1 | This has been shown to facilitate inferring 3D human pose and shape from images via different learning models. | background 1.00 | 0.05 | 0.02 | asserted 1.00 |
| s2 | However, not all pose and shape parameter values yield physically-plausible or even realistic body meshes. | background 0.70 | 0.11 | 0.03 | asserted 0.96 |
| s3 | 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. | background 0.64 | 0.10 | 0.03 | hedged possibility 0.92 |
| s4 | In this paper, we therefore learn a prior that restricts the SMPL parameters to values that produce realistic poses via adversarial training. | contribution 0.51 | 0.92 | 0.19 | asserted 0.98 |
| s5 | 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 when used for regression from images. | result 0.97 | 0.96 | 0.66 | asserted 1.00 |
| s6 | We found that the prior based on spherical distribution gets the best results. | result 1.00 | 0.95 | 0.13 | asserted 1.00 |
| s7 | Furthermore, in all these tasks, it outperforms the state-of-the-art VAE-based approach to constraining the SMPL parameters. | result 1.00 | 0.91 | 0.22 | asserted 1.00 |
What was examined, and what was not
- arXiv listing abstract v1
- full text
- figures and tables
- references and citing sentences
- relations to other papers
- P abstract
- S sections
- T body sentences
- C citations
- R cross-paper relations
The paper
Original 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 when used for regression from images. We found that the prior based on spherical distribution gets the best results. Furthermore, in all these tasks, it outperforms the state-of-the-art VAE-based approach to constraining the SMPL parameters.
Human review
How the literature uses it
1 arXiv paper cite this in 1 sentence
What it cites · 2 sentences without a confident label
Uses the method or tool of 1 paper
-
To demonstrate this, we start from the model of [ref] , whose architecture is a Resnet50 model based on the one of [ref] .
Exemplar Fine-Tuning for 3D Human Model Fitting Towards In-the-Wild 3D Human Pose Estimation · Using the Generator as a Universal Prior · 0.80
Uses data or values from 1 paper
-
It is pretrained on pseudo ground-truth COCO [ref] dataset obtained by [ref] .
Exemplar Fine-Tuning for 3D Human Model Fitting Towards In-the-Wild 3D Human Pose Estimation · Using the Generator as a Universal Prior · 0.84
Compares with 1 paper
-
Our models also report better accuracy compared to [ref] , for which the accuracy of the pre-trained model, provided by the authors, is reported.
Exemplar Fine-Tuning for 3D Human Model Fitting Towards In-the-Wild 3D Human Pose Estimation · Image to Mesh Regression · 0.95
Cites as background 1 paper
-
GAN-based approaches [ref] have used several types of distributions from which to draw their latent vectors, including Gaussian, Uniform, and Spherical distributions.
CAMPARI: Camera-Aware Decomposed Generative Neural Radiance Fields · Distribution over Latent Vectors · 0.96
Discussion
Know something about this paper? Leave a comment, or highlight any passage above to comment on it.