Extending the Perceptual User Interface to Recognise Movement

Abstract

AbstractPerceptual User Interfaces (PUIs) automatically extract user input from natural and implicit components of human activity such as gestures, direction of gaze, facial expression and body movement. This paper presents a Continuous Human Movement Recognition (CHMR) system for recognising a large range of specific movement skills from continuous 3D full-body motion. A new methodology defines an alphabet of dynemes, units of full-body movement skills, to enable recognition of diverse skills. Using multiple Hidden Markov Models, the CHMR system attempts to infer the movement skill that could have produced the observed sequence of dynemes. This approach enables the CHMR system to track and recognise hundreds of full-body movement skills from gait to twisting summersaults. This extends the perceptual user interface beyond frontal posing or only tracking one hand to recognise and understand full-body movement in terms of everyday activities.KeywordsHide Markov ModelMotion VectorContext ModelAmerican SignSkill ModelThese keywords were added by machine and not by the authors. This process is experimental and the keywords may be updated as the learning algorithm improves.