Controllable ECG Synthesis Through Neural Network Input Modelling

Detalhes bibliográficos
Autor(a) principal: Herrera, Ana Laura Batista
Data de Publicação: 2023
Tipo de documento: Dissertação
Idioma: eng
Título da fonte: Repositório Científico de Acesso Aberto de Portugal (Repositórios Cientìficos)
Texto Completo: http://hdl.handle.net/10362/162739
Resumo: Artificial Intelligence (AI) has brought promising innovations to healthcare, particularly in patient diagnosis and monitoring. However, challenges such as data scarcity, quality, and privacy concerns often hinder its effective implementation. Electrocardiogram (ECG) signals are crucial for cardiac disease diagnosis, but collecting these signals can be challenging due to patient privacy concerns and the rarity of some cardiac conditions. This thesis presents a novel method for synthesising ECG signals in a controllable manner, with the aim of augmenting existing datasets and, by extension, promoting the development of more ethically responsible AI algorithms in healthcare. The methods used in this thesis include the usage of a non-linear modified sawtooth signal as an input to train a Gated Recurrent Unit (GRU) generator. Then, by modifying this input signal, with alterations that result in parameterised changes to the synthetic signal, it was possible to generate different ECG traces, such as normal sinus rhythms, tachycardia, bradycardia, and first and second-degree atrioventricular (AV) blocks. The model successfully retained the specific ECG characteristics of individual subjects across these rhythms and arrhythmias. In addition, a Random Forest (RF) model was trained and then tested to classify the synthetic arrhythmias. Once this was successfully achieved, the synthetic arrhythmias were used to augment the previous training dataset. The results showed improved performance on the same test set, highlighting the utility and effectiveness of the synthetic signals. This thesis contributes to the field of AI in healthcare by developing new methods to generate data for underrepresented groups and providing evaluation metrics to assess the quality of synthetic data. Ultimately, this work serves as a foundation for the equitable use of synthetic signals in healthcare AI, promoting unbiased and ethically responsible applications.
id RCAP_fae9d18e2760ce17829bae586ac40ec0
oai_identifier_str oai:run.unl.pt:10362/162739
network_acronym_str RCAP
network_name_str Repositório Científico de Acesso Aberto de Portugal (Repositórios Cientìficos)
repository_id_str 7160
spelling Controllable ECG Synthesis Through Neural Network Input ModellingECG synthesiscontrollable synthesisGRUdata augmentationarrhythmia synthesisDomínio/Área Científica::Engenharia e Tecnologia::Outras Engenharias e TecnologiasArtificial Intelligence (AI) has brought promising innovations to healthcare, particularly in patient diagnosis and monitoring. However, challenges such as data scarcity, quality, and privacy concerns often hinder its effective implementation. Electrocardiogram (ECG) signals are crucial for cardiac disease diagnosis, but collecting these signals can be challenging due to patient privacy concerns and the rarity of some cardiac conditions. This thesis presents a novel method for synthesising ECG signals in a controllable manner, with the aim of augmenting existing datasets and, by extension, promoting the development of more ethically responsible AI algorithms in healthcare. The methods used in this thesis include the usage of a non-linear modified sawtooth signal as an input to train a Gated Recurrent Unit (GRU) generator. Then, by modifying this input signal, with alterations that result in parameterised changes to the synthetic signal, it was possible to generate different ECG traces, such as normal sinus rhythms, tachycardia, bradycardia, and first and second-degree atrioventricular (AV) blocks. The model successfully retained the specific ECG characteristics of individual subjects across these rhythms and arrhythmias. In addition, a Random Forest (RF) model was trained and then tested to classify the synthetic arrhythmias. Once this was successfully achieved, the synthetic arrhythmias were used to augment the previous training dataset. The results showed improved performance on the same test set, highlighting the utility and effectiveness of the synthetic signals. This thesis contributes to the field of AI in healthcare by developing new methods to generate data for underrepresented groups and providing evaluation metrics to assess the quality of synthetic data. Ultimately, this work serves as a foundation for the equitable use of synthetic signals in healthcare AI, promoting unbiased and ethically responsible applications.A Inteligência Artificial (IA) trouxe inovações promissoras para os cuidados de saúde, nomeadamente no diagnóstico e monitorização dos doentes. No entanto, desafios como a escassez de dados, a qualidade e as preocupações com a privacidade impedem frequentemente a utilização da IA nesta área. Os sinais de eletrocardiograma (ECG) são cruciais para o diagnóstico de doenças cardíacas, mas a recolha destes sinais pode ser um desafio devido às preocupações com a privacidade do paciente e à raridade de algumas doenças cardíacas. Esta tese apresenta um novo método para sintetizar sinais de ECG de uma forma controlável, com o objetivo de aumentar os conjuntos de dados existentes e promover o desenvolvimento de algoritmos de IA mais eticamente responsáveis na área da saúde. Os métodos utilizados nesta tese incluem a utilização de um sinal dente de serra modificado como entrada para treinar uma Gated Recurrent Unit (GRU) e, em seguida, a utilização de diferentes versões deste sinal de entrada, com alterações que resultam em mudanças parametrizadas do sinal sintético. Isto permite a geração de diferentes ritmos de ECG, tais como ritmos sinusais normais, taquicardia, bradicardia e vários bloqueios atrioventriculares (AV). O modelo reteve com êxito as características específicas do ECG de sujeitos em todos os ritmos e arritmias. Além disso, um modelo Random Forest (RF) foi treinado e depois testado para classificar as arritmias sintéticas. Posteriormente, as arritmias sintéticas foram usadas para aumentar o conjunto de dados de treino anterior, e os testes preliminares mostraram um melhor desempenho no mesmo conjunto de teste, destacando a utilidade e a eficácia dos sinais sintetizados. Esta tese contribui para a inovação na área da IA para a saúde, desenvolvendo novos métodos para gerar dados para grupos sub-representados e fornecendo métricas para avaliação da qualidade dos dados sintéticos. Em última análise, este trabalho serve de base para a utilização equitativa de dados sintéticos na IA aplicada à medicina, garantindo modelos imparciais e eticamente responsáveis.Gamboa, HugoRUNHerrera, Ana Laura Batista2024-01-25T10:37:28Z2023-112023-11-01T00:00:00Zinfo:eu-repo/semantics/publishedVersioninfo:eu-repo/semantics/masterThesisapplication/pdfhttp://hdl.handle.net/10362/162739enginfo:eu-repo/semantics/openAccessreponame:Repositório Científico de Acesso Aberto de Portugal (Repositórios Cientìficos)instname:Agência para a Sociedade do Conhecimento (UMIC) - FCT - Sociedade da Informaçãoinstacron:RCAAP2024-03-11T05:45:43Zoai:run.unl.pt:10362/162739Portal AgregadorONGhttps://www.rcaap.pt/oai/openaireopendoar:71602024-03-20T03:59:03.515899Repositório Científico de Acesso Aberto de Portugal (Repositórios Cientìficos) - Agência para a Sociedade do Conhecimento (UMIC) - FCT - Sociedade da Informaçãofalse
dc.title.none.fl_str_mv Controllable ECG Synthesis Through Neural Network Input Modelling
title Controllable ECG Synthesis Through Neural Network Input Modelling
spellingShingle Controllable ECG Synthesis Through Neural Network Input Modelling
Herrera, Ana Laura Batista
ECG synthesis
controllable synthesis
GRU
data augmentation
arrhythmia synthesis
Domínio/Área Científica::Engenharia e Tecnologia::Outras Engenharias e Tecnologias
title_short Controllable ECG Synthesis Through Neural Network Input Modelling
title_full Controllable ECG Synthesis Through Neural Network Input Modelling
title_fullStr Controllable ECG Synthesis Through Neural Network Input Modelling
title_full_unstemmed Controllable ECG Synthesis Through Neural Network Input Modelling
title_sort Controllable ECG Synthesis Through Neural Network Input Modelling
author Herrera, Ana Laura Batista
author_facet Herrera, Ana Laura Batista
author_role author
dc.contributor.none.fl_str_mv Gamboa, Hugo
RUN
dc.contributor.author.fl_str_mv Herrera, Ana Laura Batista
dc.subject.por.fl_str_mv ECG synthesis
controllable synthesis
GRU
data augmentation
arrhythmia synthesis
Domínio/Área Científica::Engenharia e Tecnologia::Outras Engenharias e Tecnologias
topic ECG synthesis
controllable synthesis
GRU
data augmentation
arrhythmia synthesis
Domínio/Área Científica::Engenharia e Tecnologia::Outras Engenharias e Tecnologias
description Artificial Intelligence (AI) has brought promising innovations to healthcare, particularly in patient diagnosis and monitoring. However, challenges such as data scarcity, quality, and privacy concerns often hinder its effective implementation. Electrocardiogram (ECG) signals are crucial for cardiac disease diagnosis, but collecting these signals can be challenging due to patient privacy concerns and the rarity of some cardiac conditions. This thesis presents a novel method for synthesising ECG signals in a controllable manner, with the aim of augmenting existing datasets and, by extension, promoting the development of more ethically responsible AI algorithms in healthcare. The methods used in this thesis include the usage of a non-linear modified sawtooth signal as an input to train a Gated Recurrent Unit (GRU) generator. Then, by modifying this input signal, with alterations that result in parameterised changes to the synthetic signal, it was possible to generate different ECG traces, such as normal sinus rhythms, tachycardia, bradycardia, and first and second-degree atrioventricular (AV) blocks. The model successfully retained the specific ECG characteristics of individual subjects across these rhythms and arrhythmias. In addition, a Random Forest (RF) model was trained and then tested to classify the synthetic arrhythmias. Once this was successfully achieved, the synthetic arrhythmias were used to augment the previous training dataset. The results showed improved performance on the same test set, highlighting the utility and effectiveness of the synthetic signals. This thesis contributes to the field of AI in healthcare by developing new methods to generate data for underrepresented groups and providing evaluation metrics to assess the quality of synthetic data. Ultimately, this work serves as a foundation for the equitable use of synthetic signals in healthcare AI, promoting unbiased and ethically responsible applications.
publishDate 2023
dc.date.none.fl_str_mv 2023-11
2023-11-01T00:00:00Z
2024-01-25T10:37:28Z
dc.type.status.fl_str_mv info:eu-repo/semantics/publishedVersion
dc.type.driver.fl_str_mv info:eu-repo/semantics/masterThesis
format masterThesis
status_str publishedVersion
dc.identifier.uri.fl_str_mv http://hdl.handle.net/10362/162739
url http://hdl.handle.net/10362/162739
dc.language.iso.fl_str_mv eng
language eng
dc.rights.driver.fl_str_mv info:eu-repo/semantics/openAccess
eu_rights_str_mv openAccess
dc.format.none.fl_str_mv application/pdf
dc.source.none.fl_str_mv reponame:Repositório Científico de Acesso Aberto de Portugal (Repositórios Cientìficos)
instname:Agência para a Sociedade do Conhecimento (UMIC) - FCT - Sociedade da Informação
instacron:RCAAP
instname_str Agência para a Sociedade do Conhecimento (UMIC) - FCT - Sociedade da Informação
instacron_str RCAAP
institution RCAAP
reponame_str Repositório Científico de Acesso Aberto de Portugal (Repositórios Cientìficos)
collection Repositório Científico de Acesso Aberto de Portugal (Repositórios Cientìficos)
repository.name.fl_str_mv Repositório Científico de Acesso Aberto de Portugal (Repositórios Cientìficos) - Agência para a Sociedade do Conhecimento (UMIC) - FCT - Sociedade da Informação
repository.mail.fl_str_mv
_version_ 1799138170932035584