Towards cross-modal comparison of human motion data

Authors:
Thomas Helten;Meinard Müller;Jochen Tautges;Andreas Weber;Hans-Peter Seidel
Affiliations:
Max-Planck-Institut für Informatik, Saarbrücken, Germany;Max-Planck-Institut für Informatik, Saarbrücken, Germany;Bonn University, Institut für Informatik II, Bonn, Germany;Bonn University, Institut für Informatik II, Bonn, Germany;Max-Planck-Institut für Informatik, Saarbrücken, Germany
Venue:
DAGM'11 Proceedings of the 33rd international conference on Pattern recognition
Year:
2011

Citing 6
Cited 0

3-D character animation using motion capture

Interactive computer animation
Animating rotation with quaternion curves

SIGGRAPH '85 Proceedings of the 12th annual conference on Computer graphics and interactive techniques
Real-Time Motion Capture for a Human Body using Accelerometers

Robotica
Motion templates for automatic classification and retrieval of motion capture data

Proceedings of the 2006 ACM SIGGRAPH/Eurographics symposium on Computer animation
Action capture with accelerometers

Proceedings of the 2008 ACM SIGGRAPH/Eurographics Symposium on Computer Animation
Motion reconstruction using sparse accelerometer data

ACM Transactions on Graphics (TOG)

Quantified Score

Hi-index	0.00

Visualization

Abstract

Analyzing human motion data has become an important strand of research in many fields such as computer animation, sport sciences, and medicine. In this paper, we discuss various motion representations that originate from different sensor modalities and investigate their discriminative power in the context of motion identification and retrieval scenarios. As one main contribution, we introduce various mid-level motion representations that allow for comparing motion data in a cross-modal fashion. In particular, we show that certain low-dimensional feature representations derived from inertial sensors are suited for specifying high-dimensional motion data. Our evaluation shows that features based on directional information outperform purely acceleration based features in the context of motion retrieval scenarios.