Gespeichert in:
Bibliographische Detailangaben
Hauptverfasser: Warner, Nikolai, Zhang, Wenjin, Badiozamani, Hamid, Essa, Irfan, Sadhwani, Apaar
Format: Preprint
Veröffentlicht: 2025
Schlagworte:
Online-Zugang:https://arxiv.org/abs/2508.07112
Tags: Tag hinzufügen
Keine Tags, Fügen Sie den ersten Tag hinzu!
_version_ 1866911574123347968
author Warner, Nikolai
Zhang, Wenjin
Badiozamani, Hamid
Essa, Irfan
Sadhwani, Apaar
author_facet Warner, Nikolai
Zhang, Wenjin
Badiozamani, Hamid
Essa, Irfan
Sadhwani, Apaar
contents Lifting-based 3D human pose estimation infers 3D joints from 2D keypoints but generalizes poorly because $(x,y)$ coordinates alone are an ill-posed, sparse representation that discards geometric information modern foundation models can recover. We propose \emph{AugLift}, which changes the representation format of lifting from 2D coordinates to a 6D geometric descriptor via two modules: (1) an \emph{Uncertainty-Aware Depth Descriptor} (UADD) -- a compact tuple $(c, d, d_{\min}, d_{\max})$ extracted from a confidence-scaled neighborhood of an off-the-shelf monocular depth map -- and (2) a scale normalization component that handles train/test distance shifts. AugLift requires no new sensors, no new data collection, and no architectural changes beyond widening the input layer; because it operates at the representation level, it is composable with any lifting architecture or domain generalization technique. In the detection setting, AugLift reduces cross-dataset MPJPE by $10.1$% on average across four datasets and four lifting architectures while improving in-distribution accuracy by $4.0$%; post-hoc analysis shows gains concentrate on novel poses and occluded joints. In the ground-truth 2D setting, combining AugLift with PoseAug's differentiable domain generalization achieves state-of-the-art cross-dataset performance ($62.4$\,mm on 3DHP, $92.6$\,mm on 3DPW; $14.5$% and $22.2$% over PoseAug), demonstrating that foundation-model depth provides genuine geometric signal complementary to explicit 3D augmentation. Code will be made publicly available.
format Preprint
id arxiv_https___arxiv_org_abs_2508_07112
institution arXiv
publishDate 2025
record_format arxiv
spellingShingle AugLift: Depth-Aware Input Reparameterization Improves Domain Generalization in 2D-to-3D Pose Lifting
Warner, Nikolai
Zhang, Wenjin
Badiozamani, Hamid
Essa, Irfan
Sadhwani, Apaar
Computer Vision and Pattern Recognition
Machine Learning
Lifting-based 3D human pose estimation infers 3D joints from 2D keypoints but generalizes poorly because $(x,y)$ coordinates alone are an ill-posed, sparse representation that discards geometric information modern foundation models can recover. We propose \emph{AugLift}, which changes the representation format of lifting from 2D coordinates to a 6D geometric descriptor via two modules: (1) an \emph{Uncertainty-Aware Depth Descriptor} (UADD) -- a compact tuple $(c, d, d_{\min}, d_{\max})$ extracted from a confidence-scaled neighborhood of an off-the-shelf monocular depth map -- and (2) a scale normalization component that handles train/test distance shifts. AugLift requires no new sensors, no new data collection, and no architectural changes beyond widening the input layer; because it operates at the representation level, it is composable with any lifting architecture or domain generalization technique. In the detection setting, AugLift reduces cross-dataset MPJPE by $10.1$% on average across four datasets and four lifting architectures while improving in-distribution accuracy by $4.0$%; post-hoc analysis shows gains concentrate on novel poses and occluded joints. In the ground-truth 2D setting, combining AugLift with PoseAug's differentiable domain generalization achieves state-of-the-art cross-dataset performance ($62.4$\,mm on 3DHP, $92.6$\,mm on 3DPW; $14.5$% and $22.2$% over PoseAug), demonstrating that foundation-model depth provides genuine geometric signal complementary to explicit 3D augmentation. Code will be made publicly available.
title AugLift: Depth-Aware Input Reparameterization Improves Domain Generalization in 2D-to-3D Pose Lifting
topic Computer Vision and Pattern Recognition
Machine Learning
url https://arxiv.org/abs/2508.07112