Dynamic Multiscale Graph Neural Networks for 3D Skeleton-Based Human Motion Prediction

TL;DR

This paper introduces a novel dynamic multiscale graph neural network (DMGNN) for predicting 3D human motion from skeleton data, leveraging adaptive multiscale graphs for improved feature learning and prediction accuracy.

Contribution

The paper proposes a new DMGNN model with multiscale graphs and a graph-based gating mechanism, enhancing motion prediction and interpretability over existing methods.

Findings

Outperforms state-of-the-art in short and long-term predictions

Effective multiscale graph modeling improves feature learning

Model is action-category-agnostic and interpretable

Abstract

We propose novel dynamic multiscale graph neural networks (DMGNN) to predict 3D skeleton-based human motions. The core idea of DMGNN is to use a multiscale graph to comprehensively model the internal relations of a human body for motion feature learning. This multiscale graph is adaptive during training and dynamic across network layers. Based on this graph, we propose a multiscale graph computational unit (MGCU) to extract features at individual scales and fuse features across scales. The entire model is action-category-agnostic and follows an encoder-decoder framework. The encoder consists of a sequence of MGCUs to learn motion features. The decoder uses a proposed graph-based gate recurrent unit to generate future poses. Extensive experiments show that the proposed DMGNN outperforms state-of-the-art methods in both short and long-term predictions on the datasets of Human 3.6M and CMU Mocap. We further investigate the learned multiscale graphs for the interpretability. The codes could be downloaded from https://github.com/limaosen0/DMGNN.