AN AN APPROACH TO HUMAN ACTIVITY CLUSTERING USING INERTIAL MEASUREMENT DATA

  • Vojkan Nikolić University of Criminal Investigation and Police Studies logoUniversity of Criminal Investigation and Police Studies
  • Milan Gnjatović
  • Dušan Joksimović
  • Nemanja Maček
  • Nebojša Budimirović
Keywords: human activity clustering, inertial measurement data, adapted Levenshtein distance, graph-based clustering

Abstract

Automatic human activity recognition is regarded as an important task in security, military and police applications. This paper reports on a pilot study of an approach to human activity clustering using inertial measurement data. At the signal level, we particularly consider the angular velocity and instantaneous acceleration data obtained from a three-axis inertial measurement unit placed on the right arm of the human subject. At the methodology level, the approach consists of three components: symbol-based modeling of spatiotemporal signals, an adaptation of the Levenshtein distance, and a graph-based clustering algorithm. A prototype system implementing the proposed approach is evaluated on recordings of six human subjects involved in four task-oriented activities with significant intercluster similarity. The clustering results are assessed using the Rand index (RI=0.921), the precision rate (P=1), the recall rate (R=0.636), and the balanced F-measure (F=0.778).

References

[1] Avci, A., Bosch, S., Marin-Perianu, M., Marin-Perianu, R., Havinga, P. (2010) Activity Recognition Using Inertial Sensing for Healthcare, Wellbeing and Sports Applications: A Survey, 23rd International Conference on Architecture of Computing Systems 2010, Hannover, Germany, pp. 1-10.
[2] De la Torre, F., Hodgins, J., Montano, J., Valcarcel, S., Forcada, R., Macey, J. (2009) Guide to the Carnegie Mellon University Multimodal Activity (CMU-MMAC) Database, Tech. report CMU-RI-TR-08-22, Robotics Institute, Carnegie Mellon University.
[3] Felzenszwalb, P.F., Huttenlocher, D.P. (2004) Efficient Graph-Based Image Segmentation, International Journal of Computer Vision, 59, pp. 167–181.
[4] Hussain, Z., Sheng, M., Zhang, W.E. (2019) Different Approaches for Human Activity Recognition: A Survey, arXiv, 1906.05074, Downloaded July 16th 2020, https://arxiv.org/abs/1906.05074.
[5] Jurafsky, D., Martin, J.H. (2009) Speech and Language Processing: An Introduction to Natural Language Processing, Speech Recognition, and Computational Linguistics, 2nd edition, Prentice-Hall.
[6] Jobanputra, C., Bavishi, J., Doshi, N. (2019) Human Activity Recognition: A Survey, Procedia Computer Science, 15, pp. 698-703.
[7] Levenshtein, V.I. (1966) Binary codes capable of correcting deletions, insertions, and reversals, Cybernetics and Control Theory, 10(8), pp. 707-710 (Original in Doklady Akademii Nauk SSSR 163(4): 845-848, 1965)
[8] Schimke, S., Vielhauer, C., Dittmann, J. (2004) Using adapted Levenshtein distance for on-line signature authentication, Proceedings of the 17th International Conference on Pattern Recognition, 2004. ICPR 2004, Cambridge, 2004, pp. 931-934, Vol.2.
Published
2020-11-27
Section
Effects of Physical Activity on Anthropological Status in Security Agency Per