Study of Generative Adversarial Networks for Generating Synthetic Data and its Application on Optoacoustic Tomography

This work proposes the use of a Generative Adversarial Network (GAN) to perform data augmentation with the goal of improving image reconstruction in Optoacustic Tomography (OAT) applications. We employ the FastGAN model, a compact net capable of generating high resolution images from small datasets....

Descripción completa

Guardado en:
Detalles Bibliográficos
Autores principales: Scopa Lopina, Alejandro, González, Martín Germán, Vera, Matías
Formato: Artículo publishedVersion
Lenguaje:Español
Publicado: FIUBA 2023
Materias:
Optoacoustic Tomography
Deep Learning
Generative Adversarial Networks
Synthetic Data
Tomografía optoacústica
Aprendizaje profundo
Redes generativas de confrontación
Datos sintéticos
Acceso en línea:https://elektron.fi.uba.ar/elektron/article/view/185
https://repositoriouba.sisbi.uba.ar/gsdl/cgi-bin/library.cgi?a=d&c=elektron&d=185_oai
Aporte de:
Repositorio Digital de la Universidad de Buenos Aires (UBA) de Universidad de Buenos Aires
id I28-R145-185_oai
record_format dspace
spelling I28-R145-185_oai2026-02-11 Scopa Lopina, Alejandro González, Martín Germán Vera, Matías 2023-12-15 This work proposes the use of a Generative Adversarial Network (GAN) to perform data augmentation with the goal of improving image reconstruction in Optoacustic Tomography (OAT) applications. We employ the FastGAN model, a compact net capable of generating high resolution images from small datasets. The quality of the generated data was assessed by two methods. First, the Fréchet distance (FID) was measured, observing a decreasing trend throughout the entire GAN training. Then, a U-Net neural network designed for a OAT system with and without augmented data was trained. In this case, the model trained with the extra data generated by the GAN achieved an appreciable improvement in the figures of merit associated with the reconstruction. En este trabajo se propone el uso de una red generativa de confrontación (GAN) para efectuar un aumento de datos con el objetivo de mejorar la reconstrucción de imágenes en sistemas para tomografía optoacústica (TOA). Se utilizó el modelo denominado FastGAN que es una red compacta, capaz de generar imágenes de alta resolución a partir de un conjunto de datos reducidos. La calidad de los datos generados se evaluó a través de dos métodos. Por un lado, se usó la distancia de inicio de Fréchet (FID), observándose una tendencia decreciente a largo de todo el entrenamiento de la GAN. En el segundo método se entrenó una red neuronal U-Net diseñada para un sistema de TOA con y sin datos aumentados. En este caso, el modelo entrenado con los datos extras aportados por la GAN logró una mejora apreciable en las figuras de mérito asociadas a la reconstrucción. application/pdf text/html https://elektron.fi.uba.ar/elektron/article/view/185 10.37537/rev.elektron.7.2.185.2023 spa FIUBA https://elektron.fi.uba.ar/elektron/article/view/185/335 https://elektron.fi.uba.ar/elektron/article/view/185/342 Derechos de autor 2023 Alejandro Scopa Lopina Elektron Journal; Vol. 7 No. 2 (2023); 61-70 Revista Elektron; Vol. 7 Núm. 2 (2023); 61-70 Revista Elektron; v. 7 n. 2 (2023); 61-70 2525-0159 2525-0159 Optoacoustic Tomography Deep Learning Generative Adversarial Networks Synthetic Data Tomografía optoacústica Aprendizaje profundo Redes generativas de confrontación Datos sintéticos Study of Generative Adversarial Networks for Generating Synthetic Data and its Application on Optoacoustic Tomography Estudio de redes generativas de confrontación para generación de datos sintéticos y su aplicación a tomografía optoacústica 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=185_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 Optoacoustic Tomography
Deep Learning
Generative Adversarial Networks
Synthetic Data
Tomografía optoacústica
Aprendizaje profundo
Redes generativas de confrontación
Datos sintéticos
spellingShingle Optoacoustic Tomography
Deep Learning
Generative Adversarial Networks
Synthetic Data
Tomografía optoacústica
Aprendizaje profundo
Redes generativas de confrontación
Datos sintéticos
Scopa Lopina, Alejandro
González, Martín Germán
Vera, Matías
Study of Generative Adversarial Networks for Generating Synthetic Data and its Application on Optoacoustic Tomography
topic_facet Optoacoustic Tomography
Deep Learning
Generative Adversarial Networks
Synthetic Data
Tomografía optoacústica
Aprendizaje profundo
Redes generativas de confrontación
Datos sintéticos
description This work proposes the use of a Generative Adversarial Network (GAN) to perform data augmentation with the goal of improving image reconstruction in Optoacustic Tomography (OAT) applications. We employ the FastGAN model, a compact net capable of generating high resolution images from small datasets. The quality of the generated data was assessed by two methods. First, the Fréchet distance (FID) was measured, observing a decreasing trend throughout the entire GAN training. Then, a U-Net neural network designed for a OAT system with and without augmented data was trained. In this case, the model trained with the extra data generated by the GAN achieved an appreciable improvement in the figures of merit associated with the reconstruction.
format Artículo
publishedVersion
author Scopa Lopina, Alejandro
González, Martín Germán
Vera, Matías
author_facet Scopa Lopina, Alejandro
González, Martín Germán
Vera, Matías
author_sort Scopa Lopina, Alejandro
title Study of Generative Adversarial Networks for Generating Synthetic Data and its Application on Optoacoustic Tomography
title_short Study of Generative Adversarial Networks for Generating Synthetic Data and its Application on Optoacoustic Tomography
title_full Study of Generative Adversarial Networks for Generating Synthetic Data and its Application on Optoacoustic Tomography
title_fullStr Study of Generative Adversarial Networks for Generating Synthetic Data and its Application on Optoacoustic Tomography
title_full_unstemmed Study of Generative Adversarial Networks for Generating Synthetic Data and its Application on Optoacoustic Tomography
title_sort study of generative adversarial networks for generating synthetic data and its application on optoacoustic tomography
publisher FIUBA
publishDate 2023
url https://elektron.fi.uba.ar/elektron/article/view/185
https://repositoriouba.sisbi.uba.ar/gsdl/cgi-bin/library.cgi?a=d&c=elektron&d=185_oai
work_keys_str_mv AT scopalopinaalejandro studyofgenerativeadversarialnetworksforgeneratingsyntheticdataanditsapplicationonoptoacoustictomography
AT gonzalezmartingerman studyofgenerativeadversarialnetworksforgeneratingsyntheticdataanditsapplicationonoptoacoustictomography
AT veramatias studyofgenerativeadversarialnetworksforgeneratingsyntheticdataanditsapplicationonoptoacoustictomography
AT scopalopinaalejandro estudioderedesgenerativasdeconfrontacionparageneraciondedatossinteticosysuaplicacionatomografiaoptoacustica
AT gonzalezmartingerman estudioderedesgenerativasdeconfrontacionparageneraciondedatossinteticosysuaplicacionatomografiaoptoacustica
AT veramatias estudioderedesgenerativasdeconfrontacionparageneraciondedatossinteticosysuaplicacionatomografiaoptoacustica
_version_ 1859522235456290816

Ejemplares similares