Real Time Gesture Recognition System to Simulate Guitar Chords

Abstract

We have all been used to buying an instrument in order to practice and learn it.
But with the advent of technology, such things are bound to change. Currently, one needs to
have an actual guitar to start learning to play it efficiently, and a major investment is
involved, which may not be necessary for beginners. In this paper we propose a novel way
to utilize the technology available to simulate the feel of playing chords on a guitar, with
the use of gesture recognition on a standard laptop webcam as a means for learning guitar
chords. This will involve the use of Computer Vision and Machine Learning to train models
to detect and track hands, while also using a model with a chord dataset to teach the
computer to detect the chord shape presented by the gesture. When making a correct
chord shape, the system will ideally detect the chord and play the respective sounds made
by the chord.