#Flow

Expressing movement quality

Flow was an improvised dance performance at the User in Flux workshop at the 2011 ACM CHI Conference on Human Factors in Computing Systems. Movement qualities are extracted in real time from the performer’s body using EffortDetect. EffortDetect is a real-time machine-learning system that applies Laban Movement Analysis, a rigorous framework for analyzing the human movement, to extract movement qualities from a moving body in the form of Laban Basic Efforts. It produces a dynamic stream of Laban Basic Effort qualities in real time. We extended the use of EffortDetect by designing a visualization system that uses movement quality parameters to generate an abstract visualization for use in dance performance. [Program PDF] I worked on the hardware and machine learning components of EffortDetect. This project was a collaboration with Pat Subyen (who designed and programmed the dynamic visual mapping), Kristin Carlson (the performer), Thecla Schiphorst, and Philippe Pasquier.



Comments are closed.