Spectrograms of baleen whale records synthesized from Autoenconder architectures: CAE, VAE and CAE-LSTM

In this paper, different architectures of simple convolutional networks are analyzed to generate synthetic spectrograms corresponding to baleen whales. Simplicity in these models plays an important role in the implementations of these type of networks on embedded systems. In addition, the scarcity o...

Descripción completa

Guardado en:

Detalles Bibliográficos
Autores principales:	Cabedio, María Celeste, Carnaghi, Marco
Formato:	Artículo publishedVersion
Lenguaje:	Español
Publicado:	FIUBA 2022
Materias:	Convolutional autoencoders recursive layers spectrograms underwater sound synthesis Autoencoders convolucionales Capas recursivas espectrogramas sonidos subcuáticos síntesis
Acceso en línea:	https://elektron.fi.uba.ar/elektron/article/view/167 https://repositoriouba.sisbi.uba.ar/gsdl/cgi-bin/library.cgi?a=d&c=elektron&d=167_oai
Aporte de:	Repositorio Digital de la Universidad de Buenos Aires (UBA) de Universidad de Buenos Aires

id	I28-R145-167_oai
record_format	dspace
spelling	I28-R145-167_oai2026-02-11 Cabedio, María Celeste Carnaghi, Marco 2022-12-15 In this paper, different architectures of simple convolutional networks are analyzed to generate synthetic spectrograms corresponding to baleen whales. Simplicity in these models plays an important role in the implementations of these type of networks on embedded systems. In addition, the scarcity of available data requires the generation of efficient models. With this aim in mind, simple Autoencoder architectures with a low number of as- sociated parameters are presented and trained in this paper. Then, adequate metrics are obtained and the corresponding comparison among the architecture alternatives is made. The obtained results show that the more straightforward architecture is, in turn, the most convenient. Finally, from these models, synthetic spectrograms are generated from few data samples are generated, employing a low complexity architecture and assuming a normal distribution of the latent space vectors from the training data. En este trabajo se analizan diferentes arquitecturas de redes convolucionales sencillas para generar espectrogramas sintéticos correspondientes a registros de audio de ballenas barbadas. La sencillez en el modelo juega un rol importante en las implementaciones de este tipo de redes sobre sistemas embebidos. Además, existe una necesidad de generar modelos eficientes frente a la escasez de datos disponibles para este tipo de aplicaciones. Con tal fin, se presentan arquitecturas de Autoencoders simples y de baja cantidad de parámetros asociados, se entrenan los modelos, se obtienen métricas adecuadas y se realizan las correspondientes comparaciones. Los resultados obtenidos demuestran que la arquitectura con una implementación más directa es, a su vez, la más conveniente. Finalmente, a partir de estos modelos, se generan espectrogramas sintéticos a partir de pocos datos de muestra, empleando una arquitectura de baja complejidad y asumiendo una distribución normal de los vectores reales. application/pdf text/html https://elektron.fi.uba.ar/elektron/article/view/167 10.37537/rev.elektron.6.2.167.2022 spa FIUBA https://elektron.fi.uba.ar/elektron/article/view/167/305 https://elektron.fi.uba.ar/elektron/article/view/167/317 Derechos de autor 2022 María Celeste Cabedio, Marco Carnaghi Elektron Journal; Vol. 6 No. 2 (2022); 129-134 Revista Elektron; Vol. 6 Núm. 2 (2022); 129-134 Revista Elektron; v. 6 n. 2 (2022); 129-134 2525-0159 2525-0159 Convolutional autoencoders recursive layers spectrograms underwater sound synthesis Autoencoders convolucionales Capas recursivas espectrogramas sonidos subcuáticos síntesis Spectrograms of baleen whale records synthesized from Autoenconder architectures: CAE, VAE and CAE-LSTM Espectrogramas de registros de Ballenas Barbadas sintetizados a partir de arquitecturas de Autoenconders: CAE, VAE y CAE-LSTM info:eu-repo/semantics/article info:eu-repo/semantics/publishedVersion https://repositoriouba.sisbi.uba.ar/gsdl/cgi-bin/library.cgi?a=d&c=elektron&d=167_oai
institution	Universidad de Buenos Aires
institution_str	I-28
repository_str	R-145
collection	Repositorio Digital de la Universidad de Buenos Aires (UBA)
language	Español
orig_language_str_mv	spa
topic	Convolutional autoencoders recursive layers spectrograms underwater sound synthesis Autoencoders convolucionales Capas recursivas espectrogramas sonidos subcuáticos síntesis
spellingShingle	Convolutional autoencoders recursive layers spectrograms underwater sound synthesis Autoencoders convolucionales Capas recursivas espectrogramas sonidos subcuáticos síntesis Cabedio, María Celeste Carnaghi, Marco Spectrograms of baleen whale records synthesized from Autoenconder architectures: CAE, VAE and CAE-LSTM
topic_facet	Convolutional autoencoders recursive layers spectrograms underwater sound synthesis Autoencoders convolucionales Capas recursivas espectrogramas sonidos subcuáticos síntesis
description	In this paper, different architectures of simple convolutional networks are analyzed to generate synthetic spectrograms corresponding to baleen whales. Simplicity in these models plays an important role in the implementations of these type of networks on embedded systems. In addition, the scarcity of available data requires the generation of efficient models. With this aim in mind, simple Autoencoder architectures with a low number of as- sociated parameters are presented and trained in this paper. Then, adequate metrics are obtained and the corresponding comparison among the architecture alternatives is made. The obtained results show that the more straightforward architecture is, in turn, the most convenient. Finally, from these models, synthetic spectrograms are generated from few data samples are generated, employing a low complexity architecture and assuming a normal distribution of the latent space vectors from the training data.
format	Artículo publishedVersion
author	Cabedio, María Celeste Carnaghi, Marco
author_facet	Cabedio, María Celeste Carnaghi, Marco
author_sort	Cabedio, María Celeste
title	Spectrograms of baleen whale records synthesized from Autoenconder architectures: CAE, VAE and CAE-LSTM
title_short	Spectrograms of baleen whale records synthesized from Autoenconder architectures: CAE, VAE and CAE-LSTM
title_full	Spectrograms of baleen whale records synthesized from Autoenconder architectures: CAE, VAE and CAE-LSTM
title_fullStr	Spectrograms of baleen whale records synthesized from Autoenconder architectures: CAE, VAE and CAE-LSTM
title_full_unstemmed	Spectrograms of baleen whale records synthesized from Autoenconder architectures: CAE, VAE and CAE-LSTM
title_sort	spectrograms of baleen whale records synthesized from autoenconder architectures: cae, vae and cae-lstm
publisher	FIUBA
publishDate	2022
url	https://elektron.fi.uba.ar/elektron/article/view/167 https://repositoriouba.sisbi.uba.ar/gsdl/cgi-bin/library.cgi?a=d&c=elektron&d=167_oai
work_keys_str_mv	AT cabediomariaceleste spectrogramsofbaleenwhalerecordssynthesizedfromautoenconderarchitecturescaevaeandcaelstm AT carnaghimarco spectrogramsofbaleenwhalerecordssynthesizedfromautoenconderarchitecturescaevaeandcaelstm AT cabediomariaceleste espectrogramasderegistrosdeballenasbarbadassintetizadosapartirdearquitecturasdeautoenconderscaevaeycaelstm AT carnaghimarco espectrogramasderegistrosdeballenasbarbadassintetizadosapartirdearquitecturasdeautoenconderscaevaeycaelstm
_version_	1859522232637718528

Spectrograms of baleen whale records synthesized from Autoenconder architectures: CAE, VAE and CAE-LSTM

Ejemplares similares