LoBSTr: Real-time Lower-body Pose Prediction from Sparse Upper-body Tracking Signals

TL;DR

This paper presents a real-time deep learning method that predicts lower-body pose from sparse upper-body tracking signals, improving robustness and reducing artifacts in VR motion capture.

Contribution

Introduces a GRU-based neural network that uses velocity-based input signals to accurately predict lower-body pose and contact states from sparse upper-body data in real-time.

Findings

Velocity representation improves correlation modeling between upper and lower body.

The method effectively reduces foot-skating and floating artifacts.

Achieves real-time performance on commercial VR tracking data.

Abstract

With the popularization of game and VR/AR devices, there is a growing need for capturing human motion with a sparse set of tracking data. In this paper, we introduce a deep neural-network (DNN) based method for real-time prediction of the lower-body pose only from the tracking signals of the upper-body joints. Specifically, our Gated Recurrent Unit (GRU)-based recurrent architecture predicts the lower-body pose and feet contact probability from past sequence of tracking signals of the head, hands and pelvis. A major feature of our method is that the input signal is represented with the velocity of tracking signals. We show that the velocity representation better models the correlation between the upper-body and lower-body motions and increase the robustness against the diverse scales and proportions of the user body than position-orientation representations. In addition, to remove foot-skating and floating artifacts, our network predicts feet contact state, which is used to post-process the lower-body pose with inverse kinematics to preserve the contact. Our network is lightweight so as to run in real-time applications. We show the effectiveness of our method through several quantitative evaluations against other architectures and input representations, with respect to wild tracking data obtained from commercial VR devices.