# AugLift: Depth-Aware Input Reparameterization Improves Domain Generalization in 2D-to-3D Pose Lifting

arXiv: https://arxiv.org/abs/2508.07112 · HTML: https://mlsanity.com/paper/2508.07112/

## Key facts

- Cited by 1 papers in the map (data through 14 Oct 2026).
- Builds on 19 of its 28 references that are in the map.
- First posted on arXiv 9 Aug 2025.

## Builds on (most cited first, 19 of 19)

- [Monocular 3D Human Pose Estimation In The Wild Using Improved CNN Supervision](https://mlsanity.com/paper/1611.09813/) (2016)
- [3D human pose estimation in video with temporal convolutions and semi-supervised training](https://mlsanity.com/paper/1811.11742/) (2018)
- [A simple yet effective baseline for 3d human pose estimation](https://mlsanity.com/paper/1705.03098/) (2017)
- [3D Human Pose Estimation with Spatial and Temporal Transformers](https://mlsanity.com/paper/2103.10455/) (2021)
- [Coarse-to-Fine Volumetric Prediction for Single-Image 3D Human Pose](https://mlsanity.com/paper/1611.07828/) (2016)
- [MotionBERT: A Unified Perspective on Learning Human Motion Representations](https://mlsanity.com/paper/2210.06551/) (2022)
- [Ordinal Depth Supervision for 3D Human Pose Estimation](https://mlsanity.com/paper/1805.04095/) (2018)
- [Unsupervised 3D Pose Estimation with Geometric Self-Supervision](https://mlsanity.com/paper/1904.04812/) (2019)
- [Single-Stage Multi-Person Pose Machines](https://mlsanity.com/paper/1908.09220/) (2019)
- [PoseAug: A Differentiable Pose Augmentation Framework for 3D Human Pose Estimation](https://mlsanity.com/paper/2105.02465/) (2021)
- [Unsupervised Geometry-Aware Representation for 3D Human Pose Estimation](https://mlsanity.com/paper/1804.01110/) (2018)
- [Predicting Camera Viewpoint Improves Cross-dataset Generalization for 3D Human Pose Estimation](https://mlsanity.com/paper/2004.03143/) (2020)
- [Global Adaptation meets Local Generalization: Unsupervised Domain Adaptation for 3D Human Pose Estimation](https://mlsanity.com/paper/2303.16456/) (2023)
- [A Dual-Augmentor Framework for Domain Generalization in 3D Human Pose Estimation](https://mlsanity.com/paper/2403.11310/) (2024)
- [A Single 2D Pose with Context is Worth Hundreds for 3D Human Pose Estimation](https://mlsanity.com/paper/2311.03312/) (2023)
- [DH-AUG: DH Forward Kinematics Model Driven Augmentation for 3D Human Pose Estimation](https://mlsanity.com/paper/2207.09303/) (2022)
- [Lifting by Image -- Leveraging Image Cues for Accurate 3D Human Pose Estimation](https://mlsanity.com/paper/2312.15636/) (2023)
- [PoseGU: 3D Human Pose Estimation with Novel Human Pose Generator and Unbiased Learning](https://mlsanity.com/paper/2207.03618/) (2022)
- [PoseBench3D: A Cross-Dataset Analysis Framework for 3D Human Pose Estimation via Pose Lifting Networks](https://mlsanity.com/paper/2505.10888/) (2025)

## Built on by (most cited first, 1 of 1)

- [VARPose: Flexible 2D Pose Densification via Visual Autoregressive Modeling for Enhanced 3D Lifting](https://mlsanity.com/paper/2608.02214/) (2026)

## Abstract (arXiv)

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 AugLift, which changes the representation format of lifting from 2D coordinates to a 6D geometric descriptor via two modules: (1) an Uncertainty-Aware Depth Descriptor (UADD) -- a compact tuple (c, d, dmin, dmax) 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.