This paper presents a study on automatic performer identification in expressive piano performances using Convolutional Neural Networks (CNNs) and expressive features. The study addresses the task of identifying virtuoso pianists, which has substantial implications for building dynamic musical instruments with intelligence and smart musical systems. The researchers leveraged large-scale expressive piano performance datasets and deep learning techniques, refining the scores for more accurate feature extraction. The proposed model outperforms the baseline, achieving 85.3% accuracy in a 6-way identification task. The refined dataset proved more apt for training a robust pianist identifier, contributing significantly to the field of automatic performer identification.
Publication date: 4 Oct 2023
Project Page: https://github.com/BetsyTang/PID-CNN
Paper: https://arxiv.org/pdf/2310.00699
