Toward Reliable Human Pose Forecasting with Uncertainty

TL;DR

This paper introduces a comprehensive benchmark and novel uncertainty modeling techniques for human pose forecasting, significantly improving short-term prediction accuracy and uncertainty estimation across multiple datasets.

Contribution

It develops an open-source library for pose forecasting with standardized evaluation and proposes new methods for modeling aleatoric and epistemic uncertainties.

Findings

Up to 25% improvement in short-term forecasting accuracy.

Enhanced uncertainty estimation with better model trustworthiness.

Consistent performance across multiple datasets.

Abstract

Recently, there has been an arms race of pose forecasting methods aimed at solving the spatio-temporal task of predicting a sequence of future 3D poses of a person given a sequence of past observed ones. However, the lack of unified benchmarks and limited uncertainty analysis have hindered progress in the field. To address this, we first develop an open-source library for human pose forecasting, including multiple models, supporting several datasets, and employing standardized evaluation metrics, with the aim of promoting research and moving toward a unified and consistent evaluation. Second, we devise two types of uncertainty in the problem to increase performance and convey better trust: 1) we propose a method for modeling aleatoric uncertainty by using uncertainty priors to inject knowledge about the pattern of uncertainty. This focuses the capacity of the model in the direction of more meaningful supervision while reducing the number of learned parameters and improving stability; 2) we introduce a novel approach for quantifying the epistemic uncertainty of any model through clustering and measuring the entropy of its assignments. Our experiments demonstrate up to $25\%$ improvements in forecasting at short horizons, with no loss on longer horizons on Human3.6M, AMSS, and 3DPW datasets, and better performance in uncertainty estimation. The code is available online at https://github.com/vita-epfl/UnPOSed.