# Multi-hypothesis 3D human pose estimation metrics favor miscalibrated distributions

arXiv: https://arxiv.org/abs/2210.11179 · HTML: https://mlsanity.com/paper/2210.11179/

## Key facts

- Cited by 6 papers in the map (data through 14 Oct 2026).
- Builds on 29 of its 48 references that are in the map.
- First posted on arXiv 20 Oct 2022.

## Builds on (most cited first, 29 of 29)

- [Inductive Representation Learning on Large Graphs](https://mlsanity.com/paper/1706.02216/) (2017)
- [On Calibration of Modern Neural Networks](https://mlsanity.com/paper/1706.04599/) (2017)
- [Density estimation using Real NVP](https://mlsanity.com/paper/1605.08803/) (2016)
- [Modeling Relational Data with Graph Convolutional Networks](https://mlsanity.com/paper/1703.06103/) (2017)
- [Glow: Generative Flow with Invertible 1x1 Convolutions](https://mlsanity.com/paper/1807.03039/) (2018)
- [Deep High-Resolution Representation Learning for Human Pose Estimation](https://mlsanity.com/paper/1902.09212/) (2019)
- [Monocular 3D Human Pose Estimation In The Wild Using Improved CNN Supervision](https://mlsanity.com/paper/1611.09813/) (2016)
- [A simple yet effective baseline for 3d human pose estimation](https://mlsanity.com/paper/1705.03098/) (2017)
- [DeepPose: Human Pose Estimation via Deep Neural Networks](https://mlsanity.com/paper/1312.4659/) (2013)
- [Learning to Estimate 3D Human Pose and Shape from a Single Color Image](https://mlsanity.com/paper/1805.04092/) (2018)
- [Coarse-to-Fine Volumetric Prediction for Single-Image 3D Human Pose](https://mlsanity.com/paper/1611.07828/) (2016)
- [Semantic Graph Convolutional Networks for 3D Human Pose Regression](https://mlsanity.com/paper/1904.03345/) (2019)
- [Learning Likelihoods with Conditional Normalizing Flows](https://mlsanity.com/paper/1912.00042/) (2019)
- [Probabilistic Modeling for Human Mesh Recovery](https://mlsanity.com/paper/2108.11944/) (2021)
- [MHFormer: Multi-Hypothesis Transformer for 3D Human Pose Estimation](https://mlsanity.com/paper/2111.12707/) (2021)
- [Joint Training of a Convolutional Network and a Graphical Model for Human Pose Estimation](https://mlsanity.com/paper/1406.2984/) (2014)
- [Compositional Human Pose Regression](https://mlsanity.com/paper/1704.00159/) (2017)
- [Graph Normalizing Flows](https://mlsanity.com/paper/1905.13177/) (2019)
- [Generating Multiple Hypotheses for 3D Human Pose Estimation with Mixture Density Network](https://mlsanity.com/paper/1904.05547/) (2019)
- [Distribution Calibration for Regression](https://mlsanity.com/paper/1905.06023/) (2019)
- [Probabilistic Monocular 3D Human Pose Estimation with Normalizing Flows](https://mlsanity.com/paper/2107.13788/) (2021)
- [Monocular 3D Human Pose Estimation by Generation and Ordinal Ranking](https://mlsanity.com/paper/1904.01324/) (2019)
- [CanonPose: Self-Supervised Monocular 3D Human Pose Estimation in the Wild](https://mlsanity.com/paper/2011.14679/) (2020)
- [Harvesting Multiple Views for Marker-less 3D Human Pose Annotations](https://mlsanity.com/paper/1704.04793/) (2017)
- [Deep Learning using Rectified Linear Units (ReLU)](https://mlsanity.com/paper/1803.08375/) (2018)
- [Generating Multiple Diverse Hypotheses for Human 3D Pose Consistent with 2D Joint Detections](https://mlsanity.com/paper/1702.02258/) (2017)
- [GraphMDN: Leveraging graph structure and deep learning to solve inverse problems](https://mlsanity.com/paper/2010.13668/) (2020)
- [Weakly Supervised Generative Network for Multiple 3D Human Pose Hypotheses](https://mlsanity.com/paper/2008.05770/) (2020)
- [Deep Monocular 3D Human Pose Estimation via Cascaded Dimension-Lifting](https://mlsanity.com/paper/2104.03520/) (2021)

## Built on by (most cited first, 6 of 6)

- [GFPose: Learning 3D Human Pose Prior with Gradient Fields](https://mlsanity.com/paper/2212.08641/) (2022)
- [Improved Trajectory Reconstruction for Markerless Pose Estimation](https://mlsanity.com/paper/2303.02413/) (2023)
- [On the Calibration of Human Pose Estimation](https://mlsanity.com/paper/2311.17105/) (2023)
- [Calibrated Uncertainty for Trustworthy Clinical Gait Analysis Using Probabilistic Multiview Markerless Motion Capture](https://mlsanity.com/paper/2601.22412/) (2026)
- [Platypose: Calibrated Zero-Shot Multi-Hypothesis 3D Human Motion Estimation](https://mlsanity.com/paper/2403.06164/) (2024)
- [You Only Flow Once: Calibrated and Real-Time Radar Pose Estimation with Multi-Hypothesis Normalizing Flows](https://mlsanity.com/paper/2608.09579/) (2026)

## Abstract (arXiv)

Due to depth ambiguities and occlusions, lifting 2D poses to 3D is a highly ill-posed problem. Well-calibrated distributions of possible poses can make these ambiguities explicit and preserve the resulting uncertainty for downstream tasks. This study shows that previous attempts, which account for these ambiguities via multiple hypotheses generation, produce miscalibrated distributions. We identify that miscalibration can be attributed to the use of sample-based metrics such as minMPJPE. In a series of simulations, we show that minimizing minMPJPE, as commonly done, should converge to the correct mean prediction. However, it fails to correctly capture the uncertainty, thus resulting in a miscalibrated distribution. To mitigate this problem, we propose an accurate and well-calibrated model called Conditional Graph Normalizing Flow (cGNFs). Our model is structured such that a single cGNF can estimate both conditional and marginal densities within the same model - effectively solving a zero-shot density estimation problem. We evaluate cGNF on the Human~3.6M dataset and show that cGNF provides a well-calibrated distribution estimate while being close to state-of-the-art in terms of overall minMPJPE. Furthermore, cGNF outperforms previous methods on occluded joints while it remains well-calibrated.