TitleSupport Vector Machine Learning for Gesture Signal Estimation with a Piezo Resistive Fabric Touch Surface
Publication TypeConference Paper
Year of Publication2010
AuthorsSchmeder, A, Freed, A
Conference NameNIME
Conference LocationSydney, Australia

The design of an unusually simple fabric-based touch andpressure sensor is introduced. An analysis of the rawsensor data is shown to have significant non-linearities andnon-uniform noise. Using support vector machine learningand a state-dependent adaptive filter it is demonstratedthat these problems can be overcome. The method isevaluated quantitatively using a statistical estimate of theinstantaneous rate of information transfer. The SVMregression alone is shown to improve the gesture signalinformation rate by up to 20% with zero added latency, andin combination with filtering by 40% subject to a constantlatency bound of 10 milliseconds.