The LAAS Parkour MoCap database

LAAS-CNRS, Toulouse, France

teaser-image

 

Introduction

This dataset contains 65 sequences of 5 parkour experts performing different parkour techniques: safety-vault, kong vault, pull-up and muscle-up as shown in the images below. These motions are highly dynamic and complex and require whole-body coordination.
Safety vault "Safety vault example"
Kong vault "Kong vault example"
Pull-up "Pull-up example"
Muscle-up "Muscle-up example"

 

Data acquisition

The anatomical model and the marker set are shown in the following image: Anatomical model and marker set

Whole-body 3D kinematic data were collected by 14 infra-red cameras sampling at 400 Hz (Vicon, Oxford Metrics, Oxford, UK) that recorded the motion of 48 reflective markers placed on each expert according to the recommendations of the International Society of Biomechanics (ISB).

 

Anatomical model and marker set

Two force plates (AMTI, Watertown, MA, USA) embedded into the floor and two rigid handle bar sensors (∅ 63 mm, SENSIX, Poitiers,Vienne, France) placed on a parkour tubular structure were used to record external forces at 2 kHz.

 

Data Format

Each parkour sequence is provided with an RGB video, ground-truth 3D marker positions and contact forces.

  • Each video is captured with a static camera from a single viewpoint. To protect privacy, the faces of the human subjects have been blurred.

  • Four types of contact forces are reconstructed: two contact forces exerted on left and right hand and two ground reaction forces exerted by the ground on the body parts in contact with it (namely feet). All forces are 6-dimensional spatial forces. One should note that sometimes the ground reaction forces aren’t captured - for example when the human subjects move out of the force platforms.

  • The marker positions, contact forces together with many details (frame rates, force platform corners, etc.) are saved in C3D files (see download link above). The RGB videos can be downloaded seperately.

 

Citation

If you find this dataset useful please cite the following work:

  • Galo Maldonado, François Bailly, Philippe Souères, Bruno Watier. Angular momentum regulation strategies for highly dynamic landing in Parkour. Computer Methods in Biomechanics and Biomedical Engineering, Taylor & Francis, 2017, 20 (sup1), pp.123-124.

  • Galo Maldonado. Analysis and generation of highly dynamic motions of anthropomorphic systems : application to parkour. Robotics [cs.RO]. Université Paul Sabatier - Toulouse III, 2017. English.

The anatomical model and the marker set we use are described in: