Violin Gesture Recognition Using FMCW Radars

Abstract

Bowing gestures are a key component of violin playing and can be analyzed to provide feedback on a violinist's performance. Radar systems have increasingly been used to recognize human movements but not yet in a musical context. In this study, a portable frequency-modulated continuous-wave (FMCW) radar is used to detect various violin bowing gestures. Range profiles and time-Doppler spectrograms are extracted from the raw signal data, and their unique characteristics allow for the differentiation of different bowing techniques and the recognition of incorrect bowing motions. The results of this study demonstrate the potential of radars in aiding musical instrument training.