Deep Composer Classification Using Symbolic Representation

TL;DR

This paper presents a deep learning approach for composer classification using symbolic MIDI data, achieving high accuracy on a classical music dataset by analyzing onset and note activation features.

Contribution

It introduces a novel neural network model that classifies composers based on symbolic representations derived from MIDI recordings, with a focus on onset and note activation features.

Findings

F1 score of 0.8333 on MAESTRO dataset

Effective classification of 13 classical composers

Utilizes symbolic domain features for composer identification

Abstract

In this study, we train deep neural networks to classify composer on a symbolic domain. The model takes a two-channel two-dimensional input, i.e., onset and note activations of time-pitch representation, which is converted from MIDI recordings and performs a single-label classification. On the experiments conducted on MAESTRO dataset, we report an F1 value of 0.8333 for the classification of 13~classical composers.