Arquitetura embarcável para detecção de eventos sonoros utilizando inteligência artificial
Resumo
Técnicas de Aprendizado de Máquina têm revolucionado as aplicações de monitoramento em ambientes inteligentes, seja qual for o tipo da grandeza monitorada. Este artigo propõe um pipeline de detecção de eventos baseado em sinais de áudio. Tal solução parte do problema de que o monitoramento de ambientes reais deve conter áudios de interesse e áudios desconhecidos, em sequências de sons muito variáveis, trazendo a necessidade de combinar diferentes técnicas de aprendizado em um só modelo. Particularmente, propomos a combinação técnicas de detecção de anomalia, seguida de classificadores de trechos de áudios e, posteriormente, um classificador final para sequências de eventos de interesse. Ainda, avaliamos o desempenho de tal modelo em uma plataforma embarcada. Resultados mostram a viabilidade do modelo, com acurácia geral de 93,75% sobre um dataset de teste, e tempo de predição de 0,45s em plataforma embarcada popular.
Palavras-chave:
Inteligência Artificial, Aprendizado profundo, Aprendizado de Máquina, Sistema Embarcado, Processamento Digital de Sinais, Processamento de Áudio, Classificação de eventos sonoros
Referências
Adavanne, S., Parascandolo, G., Pertil ̈a, P., Heittola, T., and Virtanen, T. (2016). Soundevent detection in multichannel audio using spatial and harmonic features. Technicalreport, DCASE2016 Challenge.
Agarap, A. F. (2018). Deep learning using rectified linear units (relu).
Chandola, V., Banerjee, A., and Kumar, V. (2009). Anomaly detection: A survey.ACMComput. Surv., 41
Chen, H., Liu, Z., Liu, Z., Zhang, P., and Yan, Y. (2019). Integrating the data augmen-tation scheme with various classifiers for acoustic scene modeling. Technical report,DCASE2019 Challenge.
Eghbal-Zadeh, H., Lehner, B., Dorfer, M., and Widmer, G. (2016). CP-JKU submissionsfor DCASE-2016: a hybrid approach using binaural i-vectors and deep convolutionalneural networks. Technical report, DCASE2016 Challenge.
Hochreiter, S. and Schmidhuber, J. (1997). Long short-term memory.Neural Comput.,9(8):1735–1780.
Kapka, S. and Lewandowski, M. (2019). Sound source detection, localization and clas-sification using consecutive ensemble of crnn models. Technical report, DCASE2019Challenge.
Koizumi, Y., Saito, S., Uematsu, H., Kawachi, Y., and Harada, N. (2019). Unsuperviseddetection of anomalous sound based on deep learning and the neyman–pearson lemma.IEEE/ACM Transactions on Audio, Speech, and Language Processing, 27(1):212–224.
Lin, L. and Wang, X. (2019). Guided learning convolution system for dcase 2019 task 4.Technical report, Institute of Computing Technology, Chinese Academy of Sciences,Beijing, China.
Mohammadi, M., Al-Fuqaha, A., Sorour, S., and Guizani, M. (2017). Deep learningfor iot big data and streaming analytics: A survey.IEEE Communications Surveys &Tutorials, PP
Munir, M., Siddiqui, S., Dengel, A., and Ahmed, S. (2018). Deepant: A deep learningapproach for unsupervised anomaly detection in time series.IEEE Access, PP:1–1
Piczak, K. J. ESC: Dataset for Environmental Sound Classification. InProceedings ofthe 23rd Annual ACM Conference on Multimedia, pages 1015–1018. ACM Press
Purwins, H., Li, B., Virtanen, T., Schluter, J., Chang, S.-Y., and Sainath, T. (2019). Deeplearning for audio signal processing.IEEE Journal of Selected Topics in Signal Pro-cessing, 13(2):206–219
Russakovsky, O., Deng, J., Su, H., Krause, J., Satheesh, S., Ma, S., Huang, Z., Karpathy,A., Khosla, A., Bernstein, M., Berg, A. C., and Fei-Fei, L. (2015). ImageNet LargeScale Visual Recognition Challenge.International Journal of Computer Vision (IJCV),115(3):211–252.
Sakashita, Y. and Aono, M. (2018). Acoustic scene classification by ensemble of spectro-grams based on adaptive temporal divisions. Technical report, DCASE2018 Challenge.
Simonyan, K. and Zisserman, A. (2014a). Very deep convolutional networks for large-scale image recognition.
Simonyan, K. and Zisserman, A. (2014b). Very deep convolutional networks for large-scale image recognition
Stowell, D., Giannoulis, D., Benetos, E., Lagrange, M., and Plumbley, M. D. (2015).Detection and classification of acoustic scenes and events.IEEE Transactions on Mul-timedia, 17(10):1733–1746
Xin, M. and Wang, Y. (2019). Research on image classification model based on deepconvolution neural network.EURASIP Journal on Image and Video Processing,2019(1):40.
Agarap, A. F. (2018). Deep learning using rectified linear units (relu).
Chandola, V., Banerjee, A., and Kumar, V. (2009). Anomaly detection: A survey.ACMComput. Surv., 41
Chen, H., Liu, Z., Liu, Z., Zhang, P., and Yan, Y. (2019). Integrating the data augmen-tation scheme with various classifiers for acoustic scene modeling. Technical report,DCASE2019 Challenge.
Eghbal-Zadeh, H., Lehner, B., Dorfer, M., and Widmer, G. (2016). CP-JKU submissionsfor DCASE-2016: a hybrid approach using binaural i-vectors and deep convolutionalneural networks. Technical report, DCASE2016 Challenge.
Hochreiter, S. and Schmidhuber, J. (1997). Long short-term memory.Neural Comput.,9(8):1735–1780.
Kapka, S. and Lewandowski, M. (2019). Sound source detection, localization and clas-sification using consecutive ensemble of crnn models. Technical report, DCASE2019Challenge.
Koizumi, Y., Saito, S., Uematsu, H., Kawachi, Y., and Harada, N. (2019). Unsuperviseddetection of anomalous sound based on deep learning and the neyman–pearson lemma.IEEE/ACM Transactions on Audio, Speech, and Language Processing, 27(1):212–224.
Lin, L. and Wang, X. (2019). Guided learning convolution system for dcase 2019 task 4.Technical report, Institute of Computing Technology, Chinese Academy of Sciences,Beijing, China.
Mohammadi, M., Al-Fuqaha, A., Sorour, S., and Guizani, M. (2017). Deep learningfor iot big data and streaming analytics: A survey.IEEE Communications Surveys &Tutorials, PP
Munir, M., Siddiqui, S., Dengel, A., and Ahmed, S. (2018). Deepant: A deep learningapproach for unsupervised anomaly detection in time series.IEEE Access, PP:1–1
Piczak, K. J. ESC: Dataset for Environmental Sound Classification. InProceedings ofthe 23rd Annual ACM Conference on Multimedia, pages 1015–1018. ACM Press
Purwins, H., Li, B., Virtanen, T., Schluter, J., Chang, S.-Y., and Sainath, T. (2019). Deeplearning for audio signal processing.IEEE Journal of Selected Topics in Signal Pro-cessing, 13(2):206–219
Russakovsky, O., Deng, J., Su, H., Krause, J., Satheesh, S., Ma, S., Huang, Z., Karpathy,A., Khosla, A., Bernstein, M., Berg, A. C., and Fei-Fei, L. (2015). ImageNet LargeScale Visual Recognition Challenge.International Journal of Computer Vision (IJCV),115(3):211–252.
Sakashita, Y. and Aono, M. (2018). Acoustic scene classification by ensemble of spectro-grams based on adaptive temporal divisions. Technical report, DCASE2018 Challenge.
Simonyan, K. and Zisserman, A. (2014a). Very deep convolutional networks for large-scale image recognition.
Simonyan, K. and Zisserman, A. (2014b). Very deep convolutional networks for large-scale image recognition
Stowell, D., Giannoulis, D., Benetos, E., Lagrange, M., and Plumbley, M. D. (2015).Detection and classification of acoustic scenes and events.IEEE Transactions on Mul-timedia, 17(10):1733–1746
Xin, M. and Wang, Y. (2019). Research on image classification model based on deepconvolution neural network.EURASIP Journal on Image and Video Processing,2019(1):40.
Publicado
30/06/2020
Como Citar
DE ARAÚJO FILHO, Luiz Carlos Silva; FIGUEIREDO, Carlos Maurício Seródio.
Arquitetura embarcável para detecção de eventos sonoros utilizando inteligência artificial. In: SIMPÓSIO BRASILEIRO DE COMPUTAÇÃO UBÍQUA E PERVASIVA (SBCUP), 12. , 2020, Cuiabá.
Anais [...].
Porto Alegre: Sociedade Brasileira de Computação,
2020
.
p. 211-220.
ISSN 2595-6183.
DOI: https://doi.org/10.5753/sbcup.2020.11227.
