Bodhidharma is the ancestor of jMIR. The features implemented in jSymbolic were originally developed as part of the Bodhidharma system, and ACE grew out of the Bodhidharma classification engine.

Bodhidharma is a unified system for performing musical genre classification of MIDI recordings, and includes a symbolic feature extractor and a machine learning-based classification engine. Bodhidharma attained the highest classification success rates in all four of the evaluated categories of the MIREX 2005 Symbolic Genre Classification Contest, the last symbolic genre classification contest held.

The Bodhidharma project included the collection of 950 labeled MIDI recordings belonging 38 different genres. These recordings and subsets of them have been used in a variety of symbolic music research projects.

Bodhidharma utilizes a sophisticated combination of flat, hierarchical and round-robin classification strategies based on classifier ensembles consisting of feedforward neural networks and k-nearest neighbour classifiers. Bodhidharma bases its classifications on 111 high-level features that it extracts from MIDI recordings. Each classifier ensemble uses genetic algorithms to evolve a weighted subset of the features that are appropriate for that particular ensemble, and genetic algorithms are also used to select amongst candidate classifier ensembles themselves. In addition to performing its essential classification tasks, Bodhidharma outputs the weightings of each feature after training, something that can have musicological implications.

Screen Shot

Related Publications

McKay, C. 2010. Automatic music classification with jMIR. Ph.D. Thesis. McGill University, Canada.

McKay, C., and I. Fujinaga. 2005. Automatic music classification and the importance of instrument identification. Proceedings of the Conference on Interdisciplinary Musicology. CD-ROM.

McKay, C., and I. Fujinaga. 2005. The Bodhidharma system and the results of the MIREX 2005 symbolic genre classification contest. Presented at the International Conference on Music Information Retrieval.

McKay, C. 2004. Automatic genre classification of MIDI recordings. M.A. Thesis. McGill University, Canada.

McKay, C. 2004. Automatic genre classification as a study of the viability of high-level features for music classification. Proceedings of the International Computer Music Conference. 367–70.

McKay, C. and I. Fujinaga. 2004. Automatic genre classification using large high-level musical feature sets. Proceedings of the International Conference on Music Information Retrieval. 525–30.

Questions and Comments


-top of page-