MMCM: Multimodality-aware Metric using Clustering-based Modes for Probabilistic Human Motion Prediction

TL;DR

This paper introduces MMCM, a new metric for probabilistic human motion prediction that evaluates diversity and validity of predicted motions by clustering motion modes, improving over existing metrics.

Contribution

The paper presents MMCM, a clustering-based metric that explicitly assesses multimodal distribution coverage and kinematic validity in probabilistic human motion prediction.

Findings

MMCM effectively identifies meaningful motion modes.

MMCM accurately scores diverse and valid motion predictions.

Clustering yields sensible mode definitions.

Abstract

This paper proposes a novel metric for Human Motion Prediction (HMP). Since a single past sequence can lead to multiple possible futures, a probabilistic HMP method predicts such multiple motions. While a single motion predicted by a deterministic method is evaluated only with the difference from its ground truth motion, multiple predicted motions should also be evaluated based on their distribution. For this evaluation, this paper focuses on the following two criteria. \textbf{(a) Coverage}: motions should be distributed among multiple motion modes to cover diverse possibilities. \textbf{(b) Validity}: motions should be kinematically valid as future motions observable from a given past motion. However, existing metrics simply appreciate widely distributed motions even if these motions are observed in a single mode and kinematically invalid. To resolve these disadvantages, this paper proposes a Multimodality-aware Metric using Clustering-based Modes (MMCM). For (a) coverage, MMCM divides a motion space into several clusters, each of which is regarded as a mode. These modes are used to explicitly evaluate whether predicted motions are distributed among multiple modes. For (b) validity, MMCM identifies valid modes by collecting possible future motions from a motion dataset. Our experiments validate that our clustering yields sensible mode definitions and that MMCM accurately scores multimodal predictions. Code: https://github.com/placerkyo/MMCM