Aprendizaje profundo y análisis de parámetros acústicos para la identificación de vocalizaciones del ganado en condiciones de confinamiento y manejo

Bryan Teixeira Paiva; Ana Paula Lüdtke Ferreira; Naylor Bastiani Perez

doi:10.24215/15146774e077

Autores/as

Bryan Teixeira Paiva Universidad Federal de Pampa, Brasil https://orcid.org/0000-0003-0031-9970
Ana Paula Lüdtke Ferreira Universidad Federal de Pampa, Brasil https://orcid.org/0000-0001-7057-9095
Naylor Bastiani Perez Universidad Federal de Pampa, Brasil https://orcid.org/0000-0002-4667-783X

DOI:

https://doi.org/10.24215/15146774e077

Palabras clave:

análisis acústico, aprendizaje profundo, galería de precisión

Resumen

Los sistemas de producción ganadera desempeñan un papel fundamental en la economía brasileña, siendo una de las principales fuentes de ingresos del país. Durante diversas etapas de la producción, como los procedimientos veterinarios, el pesaje y el transporte, los animales son sometidos a diferentes niveles de manejo, cada uno con el potencial de causarles estrés. El estrés animal impacta significativamente las propiedades de la carne, reduciendo su calidad al punto de volverla inadecuada para el consumo humano. Este estudio analizó las vocalizaciones de bovinos en dos condiciones psicológicamente distintas: confinamiento y manejo. El objetivo principal fue identificar situaciones de estrés mediante dos enfoques: el análisis de parámetros acústicos y el uso de aprendizaje profundo aplicado a los patrones sonoros emitidos por los animales en cada situación. En el análisis acústico, se realizó un estudio estadístico de los parámetros de frecuencia fundamental (F0), formantes espectrales (F1–F4), jitter, shimmer, armonía e intensidad. En el estudio de aprendizaje profundo, se implementaron tres arquitecturas de redes neuronales convolucionales (CNN), utilizando Coeficientes Cepstrales en Frecuencia Mel (MFCC) para la extracción de características acústicas. Los resultados del análisis acústico revelaron diferencias significativas (p<0,001) entre los parámetros de vocalizaciones estresadas y no estresadas en la mayoría de los casos. Por su parte, los resultados de las redes neuronales mostraron que las arquitecturas básica, intermedia y robusta lograron puntajes F1 del 96,97%, 97,90% y 98,74%, respectivamente.

Referencias

Arksey, H., O’Malley, L.: Scoping studies: Towards a methodological framework. Int. J. Social Research Methodology 8(1), 19–32 (2005)

Chelotti, J.O., Vanrell, S.R., Milone, D.H., Utsumi, S.A., Galli, J.R., Rufiner, H.L., Giovanini, L.L.: A real-time algorithm for acoustic monitoring of ingestive behavior of grazing cattle. Computers and Electronics in Agriculture 127, 64–75 (2016)

Chung, Y., Lee, J., Oh, S., Park, D., Chang, H., Kim, S.: Automatic detection of cow’s oestrus in audio surveillance system. Asian-Australasian journal of animal sciences 26(7), 1030 (2013)

Clapham, W.M., Fedders, J.M., Beeman, K., Neel, J.P.: Acoustic monitoring system to quantify ingestive behavior of free-grazing cattle. Computers and Electronics in Agriculture 76(1), 96–104 (2011)

da Costa, P., JR, M.: Ambiência na produção de bovinos de corte a pasto. Anais de Etologia 18, 26–42 (2000)

Dave, N.: Feature extraction methods LPC, PLP and MFCC in speech recognition. International Journal for Advance Research in Engineering and Technology 1(6), 1–4 (2013)

Deshmukh, O., Rajput, N., Singh, Y., Lathwal, S.: Vocalization patterns of dairy animals to detect animal state. In: Proceedings of the 21st International Conference on Pattern Recognition (ICPR2012). pp. 254–257. IEEE (2012)

Gavojdian, D., Lazebnik, T., Mincu, M., Oren, A., Nicolae, I., Zamansky, A.: Bovinetalk: Machine learning for vocalization analysis of dairy cattle under negative affective states. arXiv preprint arXiv:2307.13994 (2023)

Gerhard, D., et al.: Pitch extraction and fundamental frequency: History and current techniques. Department of Computer Science, University of Regina Regina, SK, Canada (2003)

Géron, A.: Hands-on machine learning with scikit-learn. Keras, and TensorFlow: Concepts, tools, and techniques to build intelligent systems 1 (2019)

Green, A.C.: Decoding Holstein-Friesian dairy cattle vocalisations: Applications for welfare assessment. Ph.D. thesis, University of Sydney (2020)

Hopkins, D.L., Bruce, H., Li, D.: Final report–causes and contributing factors to dark cutting: Current trends and future directions (2016)

Ikeda, Y., Ishii, Y.: Recognition of two psychological conditions of a single cow by her voice. Computers and Electronics in Agriculture 62(1), 67–72 (2008)

Jahns, G.: Call recognition to identify cow conditions—a call-recogniser translating calls to text. Computers and Electronics in Agriculture 62(1), 54–58 (2008)

Jorquera-Chavez, M., Fuentes, S., Dunshea, F.R., Jongman, E.C., Warner, R.D.: Computer vision and remote sensing to assess physiological responses of cattle to pre-slaughter stress, and its impact on beef quality: A review. Meat Science 156, 11–22 (2019)

Jung, D.H., Kim, N.Y., Moon, S.H., Jhin, C., Kim, H.J., Yang, J.S., Kim, H.S., Lee, T.S., Lee, J.Y., Park, S.H.: Deep learning-based cattle vocal classification model and real-time livestock monitoring system with noise filtering. Animals 11(2), 357 (2021)

Lee, J., Noh, B., Jang, S., Park, D., Chung, Y., Chang, H.H.: Stress detection and classification of laying hens by sound analysis. Asian-Australasian Journal of Animal Sciences 28(4), 592 (2015)

Lee, J., Zuo, S., Chung, Y., Park, D., Chang, H.H., Kim, S.: Formant-based acoustic features for cow’s estrus detection in audio surveillance system. In: 2014 11th IEEE International Conference on Advanced Video and Signal Based Surveillance (AVSS). pp. 236–240. IEEE (2014)

Linhart, P., Ratcliffe, V.F., Reby, D., Špinka, M.: Expression of emotional arousal in two different piglet call types. PloS one 10(8), e0135414 (2015)

Maigrot, A.L., Hillmann, E., Briefer, E.F.: Encoding of emotional valence in wild boar (sus scrofa) calls. Animals 8(6), 85 (2018)

Manteuffel, G., Schön, P.C.: Measuring pig welfare by automatic monitoring of stress calls. Agrartechische Berichte 29(1) (2002)

Meen, G., Schellekens, M., Slegers, M., Leenders, N., Van Erp-van der Kooij, E., Noldus, L.P.: Sound analysis in dairy cattle vocalisation as a potential welfare monitor. Computers and Electronics in Agriculture 118, 111–115 (2015)

Molento, C.F.M.: Bem-estar e produção animal: aspectos econômicos-revisão. Archives of Veterinary Science 10(1) (2005)

Moura, D., Silva, W., Naas, I., Tolón, Y., Lima, K., Vale, M.: Real time computer stress monitoring of piglets using vocalization analysis. Computers and Electronics in Agriculture 64(1), 11–18 (2008)

Paiva, B.T.: Um estudo sobre modelos de redes neurais para identificação de estresse em vocalizações de gado bovino. Master’s thesis, Programa de Pós-graduação em Computação Aplicada, Universidade Federal do Pampa (2024)

Pandeya, Y.R., Bhattarai, B., Afzaal, U., Kim, J.B., Lee, J.: A monophonic cow sound annotation tool using a semi-automatic method on audio/video data. Livestock Science 256, 104811 (2022)

Polizel Neto, A., Zanco, N., Lolatto, D.C., Moreira, P.S., Dromboski, T.: Perdas econômicas ocasionadas por lesões em carcaças de bovinos abatidos em matadourofrigorífico do norte de Mato Grosso. Pesquisa Veterinária Brasileira 35(4), 324–328 (2015)

Röttgen, V., Schön, P., Becker, F., Tuchscherer, A., Wrenzycki, C., Düpjan, S., Puppe, B.: Automatic recording of individual oestrus vocalisation in group-housed dairy cattle: development of a cattle call monitor. Animal 14(1), 198–205 (2020)

Şaşmaz, E., Tek, F.B.: Animal sound classification using a convolutional neural network. In: 2018 3rd International Conference on Computer Science and Engineering (UBMK). pp. 625–629. IEEE (2018)

Sattar, F.: A context-aware method-based cattle vocal classification for livestock monitoring in smart farm. Chemistry Proceedings 10(1) (2022). https://doi.org/10.3390/IOCAG2022-12233

Shorten, P.: Acoustic sensors for detecting cow behaviour. Smart Agricultural Technology 3, 100071 (2023). https://doi.org/https://doi.org/10.1016/j.atech.2022.100071

Silaparasetty, V.: Deep Learning Projects Using TensorFlow 2. Springer (2020)

Tiwari, V.: MFCC and its applications in speaker recognition. International Journal on Emerging Technologies 1(1), 19–22 (2010)

de la Torre, M.P., Briefer, E.F., Reader, T., McElligott, A.G.: Acoustic analysis of cattle (bos taurus) mother–offspring contact calls from a source–filter theory perspective. Applied Animal Behaviour Science 163, 58–68 (2015). https://doi.org/https://doi.org/10.1016/j.applanim.2014.11.017

Vidana-Vila, E., Malé, J., Freixes, M., Solıs-Cifré, M., Jiménez, M., Larrondo, C., Guevara, R., Miranda, J., Duboc, L., Mainau, E., et al.: Automatic detection of cow vocalizations using convolutional neural networks. Tampere, Finland (2023)

Yajuvendra, S., Lathwal, S.S., Rajput, N., Raja, T.V., Gupta, A.K., Mohanty, T.K., Ruhil, A.P., Chakravarty, A.K., Sharma, P.C., Sharma, V., et al.: Effective and accurate discrimination of individual dairy cattle through acoustic sensing. Applied Animal Behaviour Science 146(1-4), 11–18 (2013)

Yeon, S.C., Jeon, J.H., Houpt, K.A., Chang, H.H., Lee, H.C., Lee, H.J.: Acoustic features of vocalizations of korean native cows (bos taurus coreanea) in two different conditions. Applied animal behaviour science 101(1-2), 1–9 (2006)

Zheng, F., Zhang, G., Song, Z.: Comparison of different implementations of MFCC. Journal of Computer Science and Technology 16, 582–589 (2001)