Two algorithms are developed here: The first is based on hidden Markov models (HMMs) and Gaussian mixture models (GMMs). The features that were used include short time Fourier transform, loudness, and spectral sparsity. The second algorithm applies a support vector machine (SVM) on a frame-based level.

Related publications

M. Chum, A. Habshush, A. Rahman, and C. Sang, “IEEE AASP scene classification challenge using hidden Markov models and frame based classification,” 2013.
[More Details] [BIBTEX] [URL (ext.)]