RRV: A Spatiotemporal Descriptor for Rigid Body Motion Recognition

TL;DR

This paper introduces the RRV descriptor, a novel spatiotemporal feature for recognizing rigid body motions that is invariant to transformations and noise, improving accuracy in motion recognition tasks.

Contribution

The paper proposes the RRV descriptor, a new method leveraging local invariants for rigid body motion recognition, outperforming previous approaches in accuracy without added computational cost.

Findings

RRV descriptor is insensitive to noise and invariant to transformations.

It outperforms previous descriptors in recognition accuracy.

The method maintains computational efficiency.

Abstract

Motion behaviors of a rigid body can be characterized by a 6-dimensional motion trajectory, which contains position vectors of a reference point on the rigid body and rotations of this rigid body over time. This paper devises a Rotation and Relative Velocity (RRV) descriptor by exploring the local translational and rotational invariants of motion trajectories of rigid bodies, which is insensitive to noise, invariant to rigid transformation and scaling. A flexible metric is also introduced to measure the distance between two RRV descriptors. The RRV descriptor is then applied to characterize motions of a human body skeleton modeled as articulated interconnections of multiple rigid bodies. To illustrate the descriptive ability of the RRV descriptor, we explore it for different rigid body motion recognition tasks. The experimental results on benchmark datasets demonstrate that this simple RRV descriptor outperforms the previous ones regarding recognition accuracy without increasing computational cost.