Music Genre Classification: A Semi-supervised Approach

Citation

Poria S, Gelbukh A, Hussain A, Bandyopadhyay S & Howard N (2013) Music Genre Classification: A Semi-supervised Approach. In: Carrasco-Ochoa J, Martinez-Trinidad J, Rodriguez J & di Baja G (eds.) Pattern Recognition: 5th Mexican Conference, MCPR 2013, Querétaro, Mexico, June 26-29, 2013. Proceedings. Lecture Notes in Computer Science, 7914. MCPR 2013 : 5th Mexican Conference on Pattern Recognition, Queretaro, Mexico, 26.06.2013-29.06.2013. Berlin Heidelberg: Springer, pp. 254-263. http://link.springer.com/chapter/10.1007/978-3-642-38989-4_26#; https://doi.org/10.1007/978-3-642-38989-4_26

Abstract
Music genres can be seen as categorical descriptions used to classify music basing on various characteristics such as instrumentation, pitch, rhythmic structure, and harmonic contents. Automatic music genre classification is important for music retrieval in large music collections on the web. We build a classifier that learns from very few labeled examples plus a large quantity of unlabeled data, and show that our methodology outperforms existing supervised and unsupervised approaches. We also identify salient features useful for music genre classification. We achieve 97.1% accuracy of 10-way classification on real-world audio collections.