Improving Music Genre Classification from Multi-Modal Properties of Music and Genre Correlations Perspective

TL;DR

This paper introduces a novel multi-modal approach for music genre classification that leverages contrastive loss and cross-modal attention to better fuse audio and lyrics features, capturing genre correlations for improved accuracy.

Contribution

The work proposes a new multi-modal method with contrastive loss and symmetric attention, effectively modeling genre correlations for multi-label music genre classification.

Findings

Achieves state-of-the-art results on Music4All dataset.

Significantly outperforms previous multi-label classification methods.

Effectively captures genre correlations to improve classification accuracy.

Abstract

Music genre classification has been widely studied in past few years for its various applications in music information retrieval. Previous works tend to perform unsatisfactorily, since those methods only use audio content or jointly use audio content and lyrics content inefficiently. In addition, as genres normally co-occur in a music track, it is desirable to capture and model the genre correlations to improve the performance of multi-label music genre classification. To solve these issues, we present a novel multi-modal method leveraging audio-lyrics contrastive loss and two symmetric cross-modal attention, to align and fuse features from audio and lyrics. Furthermore, based on the nature of the multi-label classification, a genre correlations extraction module is presented to capture and model potential genre correlations. Extensive experiments demonstrate that our proposed method significantly surpasses other multi-label music genre classification methods and achieves state-of-the-art result on Music4All dataset.