Finger musical gesture recognition in 3D space without any tangible instrument for performing arts

This paper proposes a marker–less computer vision methodology for the simultaneous recognition of complex finger musical gestures performed in space without any tangible musical instrument. Image analysis techniques are applied in order to detect and identify the fingertips on a video. Scale and rotation invariance techniques are also applied. The finger gesture recognition and prediction is based on the stochastic modelling of the extracted high–level features with the help of hidden Markov models and Gaussian mixture models. The applications concern directly the finger gesture control of sound in performing arts. The proposed system can be considered in the long term as an intangible musical instrument that implies 'everyday gestures' and reduces the gap between non–musicians or blind people and music.