Kamrad Khoshhal Roudposhti - ISR - DEEC, University of Coimbra, Portugal
Jorge Dias

In this paper, an improved approach to analyse human body motion, based on Laban Movement Analysis (LMA), which is a known human movement descriptor, is proposed. In communication between people, the information which comes from their bodies, is indispensable. Thus we preposed a new probabilistic computational human movement description framework which can be used to analyse different body-motion based, human acivities. The contribution of this paper is to define a part of this framework, independent of environment context using components of the LMA. We attempt to use LMA parameters in fequency domain to describe human movement based on 3D features by using body-center as an origin of the reference coordinate. This approach allows us to analyse human movements parameters and features, more sufficient and reliable and with less data redundancy in LMA-based rule, as can be seen in the exprimental part

Body motion analysis, Laban movement analysis, Bayesian approach and frequency-based feature extraction