Feature Extraction and Expertise Analysis of Pianists' Motion-Captured Finger Gestures.

This paper investigates the analysis of expert piano playing gestures. It aims to extract quantitative and objective features to represent pianists’ hands gestures, and more specifically to enable characterization of the expertise level of pianists. To do so, four pianists with different expertise levels were recorded with a marker-based optical motion capture system while playing six different piano pieces. Movements were decomposed with principal component analysis, leading to uncorrelated subparts called eigenmovements. We observed that four eigenmovements allowed representation of the original movement with 80% accuracy, and less than ten eigenmovements were sufficient to represent it with 95% accuracy. The eigenvalues, representing the contribution of each eigenmovement in the original movement, allowed comparison of pianists with each other, and showed that more trained pianists seemed to use more eigenmovements, reflecting a better motor control of their hands.