Extracting users' intended nuances from their expressed movements: in quadruple movements

Abstract

We propose a method for extracting users' intended nuances from their expressed quadruple movements. Specifically, this method can quantify such nuances as a four dimensional vector representation {sharpness, softness, dynamics, largeness}. We then show an example of a music application based on this method that changes the volume of assigned music tracks in accordance with each attribute of the vector representation extracted from their quadruple movements like a music conductor.