An online EM algorithm in hidden (semi-)Markov models for audio segmentation and clustering

Alberto Bietti (1) , Francis Bach (2, 3) , Arshia Cont (4, 1)

1 MuTant - Synchronous Realtime Processing and Programming of Music Signals
2 SIERRA - Statistical Machine Learning and Parsimony
3 LIENS - Laboratoire d'informatique de l'école normale supérieure
4 Repmus - Représentations musicales

Audio segmentation is an essential problem in many audio signal processing tasks, which tries to segment an audio signal into homogeneous chunks. Rather than separately finding change points and computing similarities between segments, we focus on joint segmentation and clustering, using the framework of hidden Markov and semi-Markov models. We introduce a new incremental EM algorithm for hidden Markov models (HMMs) and show that it compares favorably to existing online EM algorithms for HMMs. We present results for real-time segmentation of musical notes and acoustic scenes.

Mots clés: online learning; audio segmentation; hidden Markov models; EM algorithm

Domaines: Machine Learning [stat.ML]; Apprentissage [cs.LG]; Son [cs.SD]

Alberto Bietti, Francis Bach, Arshia Cont. An online EM algorithm in hidden (semi-)Markov models for audio segmentation and clustering. ICASSP 2015 - 40th IEEE International Conference on Acoustics, Speech and Signal Processing, Apr 2015, Brisbane, Australia. ⟨hal-01115826⟩