Filters








7 Hits in 4.5 sec

Music structure analysis based on an LSTM-HSMM hybrid model

Go Shibata, Ryo Nishikimi, Kazuyoshi Yoshii
2020 Zenodo  
The experimental results show that the proposed LSTM-HSMM hybrid model outperformed a conventional HSMM.  ...  This paper describes a statistical music structure analysis method that splits an audio signal of popular music into musically meaningful sections at the beat level and classifies them into predefined  ...  Yoshii, "Music structure analysis based on an LSTM-HSMM hybrid model", in Proc. of the 21st Int. Society for Music Information Retrieval Conf., Montréal, Canada, 2020.  ... 
doi:10.5281/zenodo.4245362 fatcat:4mutrgwjhjdxdmxccbsu45vspu

Machine Learning for Data-Driven Movement Generation: a Review of the State of the Art [article]

Omid Alemi, Philippe Pasquier
2019 arXiv   pre-print
We cover topics such as high-level movement characterization, training data, features representation, machine learning models, and evaluation methods.  ...  Movement Analysis research, and dance and music studies.  ...  Crnkovic-Friis and Crnkovic-Friis (2016) 29 train an Long Short-Term Memory (LSTM) on 3D joint positions of a dancer.  ... 
arXiv:1903.08356v1 fatcat:wtqawbramvdx3kz6ffgp2sv3ja

Supervised Metric Learning for Music Structure Features [article]

Ju-Chiang Wang, Jordan B. L. Smith, Wei-Tsung Lu, Xuchen Song
2022 arXiv   pre-print
Music structure analysis (MSA) methods traditionally search for musically meaningful patterns in audio: homogeneity, repetition, novelty, and segment-length regularity.  ...  The trained model extracts features that can be used in existing MSA algorithms.  ...  Taking the converse approach, [2] used supervision to model how traditional features (MFCCs, CQT, etc.) relate to music structure, using an LSTM combined with a Hidden semi-Markov Model.  ... 
arXiv:2110.09000v2 fatcat:jkckpxf4tjf3lhgcvss4ukcktq

Supervised Metric Learning For Music Structure Features

Ju-Chiang Wang, Jordan B. L. Smith, Wei-Tsung Lu, Xuchen Song
2021 Zenodo  
Music structure analysis (MSA) methods traditionally search for musically meaningful patterns in audio: homogeneity, repetition, novelty, and segment-length regularity.  ...  The trained model extracts features that can be used by existing MSA algorithms.  ...  Taking the converse approach, [2] used supervision to model how traditional features (MFCCs, CQT, etc.) relate to music structure, using an LSTM combined with a Hidden semi-Markov Model.  ... 
doi:10.5281/zenodo.5624427 fatcat:cqqt5mmqzbefvcl55nbfrw245q

A Survey of Human Activity Recognition in Smart Homes Based on IoT Sensors Algorithms: Taxonomies, Challenges, and Opportunities with Deep Learning

Damien Bouchabou, Sao Mai Nguyen, Christophe Lohr, Benoit LeDuc, Ioannis Kanellos
2021 Sensors  
The hybrid structure that combines a CNN 1D and a LSTM is the best performing one. Outlines A number of algorithms have been studied for HAR in smart homes.  ...  [74] studied different neural network architectures, such as MLP, CNN, LSTM, GRU, or hybrid structures, to evaluate which structure is the most efficient.  ... 
doi:10.3390/s21186037 pmid:34577243 fatcat:vo5aqjlnbvhbtiktsamoybh2i4

Lombard Speech Synthesis Using Transfer Learning in a Tacotron Text-to-Speech System

Bajibabu Bollepalli, Lauri Juvela, Paavo Alku
2019 Interspeech 2019  
The performance of the vocoders was evaluated in the context of analysis-synthesis and text-to-speech (TTS). Here, the TTS systems were developed using an LSTM-based acoustic model.  ...  LSTM-based SPSS system that was trained on normal speaking style.  ... 
doi:10.21437/interspeech.2019-1333 dblp:conf/interspeech/BollepalliJA19 fatcat:5uz43svog5erzev5nzakdnc4qe

Intonation Modelling for Speech Synthesis and Emphasis Preservation

Pierre-Edouard Jean Charles Honnet
2017
A motivation for such a model is its theoretical language independence, based on the fact that humans share the same vocal apparatus.  ...  An automatic parameter extraction method which integrates a perceptually relevant measure is proposed with the model. This approach is evaluated and compared with the standard command-response model.  ...  as only an analysis and resynthesis is performed.  ... 
doi:10.5075/epfl-thesis-7520 fatcat:4g5qshl5mrgfbacc54yjeeqzbm