Dilated Temporal Fully-Convolutional Network for Semantic Segmentation of Motion Capture Data

TL;DR

This paper introduces an automatic semantic segmentation framework for motion capture data using a dilated temporal fully-convolutional network, improving accuracy and robustness over existing models.

Contribution

The paper presents a novel dilated temporal fully-convolutional network for automatic motion capture segmentation, outperforming state-of-the-art and other sequence models.

Findings

Outperforms state-of-the-art action segmentation models

Demonstrates robustness against noisy labels

Effective in segmenting motion capture sequences

Abstract

Semantic segmentation of motion capture sequences plays a key part in many data-driven motion synthesis frameworks. It is a preprocessing step in which long recordings of motion capture sequences are partitioned into smaller segments. Afterwards, additional methods like statistical modeling can be applied to each group of structurally-similar segments to learn an abstract motion manifold. The segmentation task however often remains a manual task, which increases the effort and cost of generating large-scale motion databases. We therefore propose an automatic framework for semantic segmentation of motion capture data using a dilated temporal fully-convolutional network. Our model outperforms a state-of-the-art model in action segmentation, as well as three networks for sequence modeling. We further show our model is robust against high noisy training labels.