Privacy-Preserving Generative Adversarial Network for Case-Based Explainability in Medical Image Analysis

Although Deep Learning models have achieved incredible results in medical image classification tasks, their lack of interpretability hinders their deployment in the clinical context. Case-based interpretability provides intuitive explanations, as it is a much more human-like approach than saliency-m...

Descripción completa

Guardado en:

Detalles Bibliográficos
Autores principales:	Helena Montenegro, Wilson Silva, Jaime S. Cardoso
Formato:	article
Lenguaje:	EN
Publicado:	IEEE 2021
Materias:	Case-based interpretability deep learning generative adversarial networks privacy-preserving machine learning medical image analysis Electrical engineering. Electronics. Nuclear engineering TK1-9971
Acceso en línea:	https://doaj.org/article/b765b6586cc548ca911a50167e55362a
Etiquetas:	Agregar Etiqueta Sin Etiquetas, Sea el primero en etiquetar este registro!

id	oai:doaj.org-article:b765b6586cc548ca911a50167e55362a
record_format	dspace
spelling	oai:doaj.org-article:b765b6586cc548ca911a50167e55362a2021-11-18T00:08:17ZPrivacy-Preserving Generative Adversarial Network for Case-Based Explainability in Medical Image Analysis2169-353610.1109/ACCESS.2021.3124844https://doaj.org/article/b765b6586cc548ca911a50167e55362a2021-01-01T00:00:00Zhttps://ieeexplore.ieee.org/document/9598877/https://doaj.org/toc/2169-3536Although Deep Learning models have achieved incredible results in medical image classification tasks, their lack of interpretability hinders their deployment in the clinical context. Case-based interpretability provides intuitive explanations, as it is a much more human-like approach than saliency-map-based interpretability. Nonetheless, since one is dealing with sensitive visual data, there is a high risk of exposing personal identity, threatening the individuals’ privacy. In this work, we propose a privacy-preserving generative adversarial network for the privatization of case-based explanations. We address the weaknesses of current privacy-preserving methods for visual data from three perspectives: realism, privacy, and explanatory value. We also introduce a counterfactual module in our Generative Adversarial Network that provides counterfactual case-based explanations in addition to standard factual explanations. Experiments were performed in a biometric and medical dataset, demonstrating the network’s potential to preserve the privacy of all subjects and keep its explanatory evidence while also maintaining a decent level of intelligibility.Helena MontenegroWilson SilvaJaime S. CardosoIEEEarticleCase-based interpretabilitydeep learninggenerative adversarial networksprivacy-preserving machine learningmedical image analysisElectrical engineering. Electronics. Nuclear engineeringTK1-9971ENIEEE Access, Vol 9, Pp 148037-148047 (2021)
institution	DOAJ
collection	DOAJ
language	EN
topic	Case-based interpretability deep learning generative adversarial networks privacy-preserving machine learning medical image analysis Electrical engineering. Electronics. Nuclear engineering TK1-9971
spellingShingle	Case-based interpretability deep learning generative adversarial networks privacy-preserving machine learning medical image analysis Electrical engineering. Electronics. Nuclear engineering TK1-9971 Helena Montenegro Wilson Silva Jaime S. Cardoso Privacy-Preserving Generative Adversarial Network for Case-Based Explainability in Medical Image Analysis
description	Although Deep Learning models have achieved incredible results in medical image classification tasks, their lack of interpretability hinders their deployment in the clinical context. Case-based interpretability provides intuitive explanations, as it is a much more human-like approach than saliency-map-based interpretability. Nonetheless, since one is dealing with sensitive visual data, there is a high risk of exposing personal identity, threatening the individuals’ privacy. In this work, we propose a privacy-preserving generative adversarial network for the privatization of case-based explanations. We address the weaknesses of current privacy-preserving methods for visual data from three perspectives: realism, privacy, and explanatory value. We also introduce a counterfactual module in our Generative Adversarial Network that provides counterfactual case-based explanations in addition to standard factual explanations. Experiments were performed in a biometric and medical dataset, demonstrating the network’s potential to preserve the privacy of all subjects and keep its explanatory evidence while also maintaining a decent level of intelligibility.
format	article
author	Helena Montenegro Wilson Silva Jaime S. Cardoso
author_facet	Helena Montenegro Wilson Silva Jaime S. Cardoso
author_sort	Helena Montenegro
title	Privacy-Preserving Generative Adversarial Network for Case-Based Explainability in Medical Image Analysis
title_short	Privacy-Preserving Generative Adversarial Network for Case-Based Explainability in Medical Image Analysis
title_full	Privacy-Preserving Generative Adversarial Network for Case-Based Explainability in Medical Image Analysis
title_fullStr	Privacy-Preserving Generative Adversarial Network for Case-Based Explainability in Medical Image Analysis
title_full_unstemmed	Privacy-Preserving Generative Adversarial Network for Case-Based Explainability in Medical Image Analysis
title_sort	privacy-preserving generative adversarial network for case-based explainability in medical image analysis
publisher	IEEE
publishDate	2021
url	https://doaj.org/article/b765b6586cc548ca911a50167e55362a
work_keys_str_mv	AT helenamontenegro privacypreservinggenerativeadversarialnetworkforcasebasedexplainabilityinmedicalimageanalysis AT wilsonsilva privacypreservinggenerativeadversarialnetworkforcasebasedexplainabilityinmedicalimageanalysis AT jaimescardoso privacypreservinggenerativeadversarialnetworkforcasebasedexplainabilityinmedicalimageanalysis
_version_	1718425247300976640

Privacy-Preserving Generative Adversarial Network for Case-Based Explainability in Medical Image Analysis

Ejemplares similares