Because a guitar playing requires a huge fingering training, it is difficult for beginners to play beautiful sound by
pressing the correct strings. This paper suggests that a guitar playing support system using image and sensor data processing can help the beginners to improve their playing skills. To recognize strings, frets, fingering, and playing sound corresponding to fingering, our system gets an initial position of the guitar and tracks the guitar position using feature points and sensor information while playing. In addition, our system detects the finger position and estimates pressed strings. In the experiments, we evaluate a success rate of the feature point matching and guitar slope detection for guitar tracking, and pressed strings estimation. Evaluations conclude that each evaluation has a high success rate.
View full abstract