Polytopes for music segmentation

Polytopic paradigms to study music, based on [1] and [2].

In near future, I should upload my own reference, detailing the specificity of both approaches of [1] and [2], which have been gathered on a same framework and on a same code project (this one).

The goal of the polytopic approach is to define a new compression criteria, then used as cost for music segmentation, based on dynamic programming. See [3] or [4] for more details on this approach.

[1] C. Guichaoua, Modèles de compression et critères de complexité pour la description et l’inférence de structure musicale. PhD thesis, 2017. [2] C. Louboutin, Modélisation multi-échelle et multi-dimensionnelle de la structure musicale par graphes polytopiques. PhD thesis, Rennes 1, 2019. [3] Sargent, G., Bimbot, F., & Vincent, E. (2016). Estimating the structural segmentation of popular music pieces under regularity constraints. IEEE/ACM Transactions on Audio, Speech, and Language Processing, 25(2), 344-358. [4] Marmoret, A., Cohen, J., Bertin, N., & Bimbot, F. (2020, October). Uncovering Audio Patterns in Music with Nonnegative Tucker Decomposition for Structural Segmentation. In ISMIR 2020-21st International Society for Music Information Retrieval.