A Guitar Chord Recognition Method Based on Depth Images of Hand Shapes

In this paper, the authors propose a recognition method of guitar chords based on depth images of hand shapes. There are many constraints to practicing a guitar such as its portability and noise. On the other hand, it is possible to practice playing guitar anytime and anywhere if its played sound or chord can be recognized from hand shape in real-time without the actual guitar. The proposed method uses machine learning and depth estimation from RGB images of left hands. The experiment results showed that the proposed method achieves higher accuracy than that of RGB images only.