Human Body Movements Classification

A. Hlaváček, V. Hanžl, and J. Havlík (Czech Republic)

Keywords

Movement Classification, Signal Processing, Supervised Learning, Hidden Markov Models

Abstract

This paper refers to the classificators which were created to serve the purpose of the classification of a motion in space, a motion of one part of the human body (a thumb in this case). Furthermore, the way of filming of the video recordings, which follow the motion of the thumb, the pre-processing of these recordings and the subsequent parametrization have been briefly mentioned. Two groups of classificators have been created, unsupervised and su pervised. The unsupervised classificators use the K-Means algorithm and were used only for labeling training data, therefore they will not be discussed in this paper. The supervised classificators are based on the use of Hidden Markov Models. Several classificators of diverse complex ity have been created for the purpose of verification of their functionality and quality.

Important Links:

Go Back