Multi-Hypothesis 3D human pose estimation metrics favor miscalibrated distributions

Paweł A. Pierzchlewicz, R. James Cotton, Mohammad Bashiri, Fabian H. Sinz

Keywords: Pose estimation, calibration, metrics, graph neural networks

TL;DR: Pose estimation metrics favor overconfident models; we propose cGNF, a model capable of maximizing likelihood and thus estimating accurate and well-calibrated distributions of 3D poses.

Abstract: 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 $\operatorname{minMPJPE}$. In a series of simulations, we show that minimizing $\operatorname{minMPJPE}$, 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 $\operatorname{minMPJPE}$. Furthermore, cGNF outperforms previous methods on occluded joints while it remains well-calibrated.

Anonymous Url: I certify that there is no URL (e.g., github page) that could be used to find authors’ identity.

No Acknowledgement Section: I certify that there is no acknowledgement section in this submission for double blind review.

Code Of Ethics: I acknowledge that I and all co-authors of this work have read and commit to adhering to the ICLR Code of Ethics

Submission Guidelines: Yes

Please Choose The Closest Area That Your Submission Falls Into: Applications (eg, speech processing, computer vision, NLP)