Gesture Unit Segmentation Using Spatial-Temporal Information and Machine Learning

Currently, automated gesture analysis is being widely used in different research areas, such as human-computer interaction or human-behavior analysis. With regard to the latter area in particular, gesture analysis is closely related to studies on human communication. Linguists and psycholinguists analyze gestures from several standpoints, and one of them is the analysis of gesture segments. The aim of this paper is to outline an approach to automate gesture unit segmentation, as a way of assisting linguistic studies. This objective was attained by employing a Machine Learning technique with the aid of a spatial-temporal data representation.