Kautz T, Groh B, Eskofier B (2015)
Publication Type: Conference contribution, Conference Contribution
Publication year: 2015
Event location: Sydney, Australia
URI: https://www.mad.tf.fau.de/files/2017/06/2015-Kautz-KDD_LSSA-SFM.pdf
The use of wearable sensors for automatic recognition of human activities has pervaded both professional and recreational sports. While many activities involving only a single athlete can be classified robustly, the automatic classification of complex activities involving several athletes is still in its infancy. In this paper, we present a novel approach for the recognition of such multi-player activities in the context of game sports. Our method is based on the fusion of position measurements with inertial measurements in a set of interaction features. We demonstrate the efficacy of our method in the recognition of tackles and scrums in Rugby Sevens. The results of our current work suggest that the proposed features can be leveraged to achieve classification accuracies of more than 97%.
APA:
Kautz, T., Groh, B., & Eskofier, B. (2015). Sensor fusion for multi-player activity recognition in game-sports. In Proceedings of the Workshop on Large-Scale Sports Analytics (21st ACM SIGKDD Conference on Knowledge Discovery and Data Mining). Sydney, Australia.
MLA:
Kautz, Thomas, Benjamin Groh, and Björn Eskofier. "Sensor fusion for multi-player activity recognition in game-sports." Proceedings of the Workshop on Large-Scale Sports Analytics (21st ACM SIGKDD Conference on Knowledge Discovery and Data Mining), Sydney, Australia 2015.
BibTeX: Download