Abstract
In this thesis, we will present a motion style synthetic method based on Laban Movement Analysis (LMA). We will study the characteristic of LMA factors and parameterize it from a set of motions with special style. The set of parameters will add special styles to the neutral motion data. We will advocate the domain of human movement observation, especially LMA and its Effort and Shape components, to archive adjustable numerical parameters and control of motion quality aspects. We will also discuss the relation between high level emotion semantics and LMA. Finally, we will implement a motion synthesizer with motion capture data driven dynamics model kernel. The user could depend on his demand, input neutral motion data and a set of emotion parameters or LMA parameters to produce the motion data and style he requires according to the parameters he sets up.