Efficient Metric Learning for the Analysis of Motion Data

TL;DR

This paper extends metric learning techniques to dynamic time warping for motion data, improving classification accuracy and interpretability by treating DTW dissimilarities as features and applying metric regularization.

Contribution

It introduces a novel approach to metric learning with DTW by leveraging component-wise dissimilarities as features and adapts metric regularization for better interpretability.

Findings

Enhanced classification accuracy on benchmark datasets

Effective transfer of metric regularization to DTW-based features

Improved interpretability of relevance profiles

Abstract

We investigate metric learning in the context of dynamic time warping (DTW), the by far most popular dissimilarity measure used for the comparison and analysis of motion capture data. While metric learning enables a problem-adapted representation of data, the majority of methods has been proposed for vectorial data only. In this contribution, we extend the popular principle offered by the large margin nearest neighbors learner (LMNN) to DTW by treating the resulting component-wise dissimilarity values as features. We demonstrate that this principle greatly enhances the classification accuracy in several benchmarks. Further, we show that recent auxiliary concepts such as metric regularization can be transferred from the vectorial case to component-wise DTW in a similar way. We illustrate that metric regularization constitutes a crucial prerequisite for the interpretation of the resulting relevance profiles.