Sensor-Based Feedback for Piano Pedagogy

摘要: Recent advances in sensor technology provide new opportunities for applications that utilize the user's movement as an input source. This thesis focuses on analysis, which is a sub-area of larger field concerned with interpretation signals human movement. Movement analysis can be used to feedback training motor tasks, interest application areas such rehabilitation, sports, and ergonomics. Consciously controllable, goal-directed movements, we call primary lead slight movements other parts body are beyond conscious control (secondary movements). Secondary generated due mechanical interaction environment physiological dependencies body. contributes methods distinguish between secondary movements. necessary order high-quality tasks show significant amount case when large because reaction forces (e.g., shooting ball) or execute task small various forms handcraft musical instrument performance). Furthermore, precise distinction check whether user keeps part still required by gymnastic exercise). Apart from sensor-based feedback, our results also improve current gesture recognition ignoring signal avoid misinterpreted execution gesture. The effectiveness proposed shown context pianist arm particularly challenging analyze. The one joint based measured particular joint, estimation key force MIDI data, joints arm. In know estimated acts, it determine hand has played note. For purpose two introduced: One method MIDI; uses data inertial sensors combination MIDI. A third Computer Vision, was originally developed sign language recognition, evaluated here tracking hands. Based methods, pedagogical were developed: supports existing piano notation checks player's conforms indicated study students music university shows potential users judge system useful technique. second visualizes allows synchronizing different performances same piece, making easy spot differences where closer examination may beneficial.