Defesa de Qualificação de Mestrado: João Gabriel Soares Ferreira
Data da publicação: 1 de novembro de 2024 Categoria: Notícias, Qualificação de Mestrado
Título: Dense Encoder ECG Filter
Data: 07/11/2024
Horário: 14h00
Local: Sala de Seminários – Bloco 952
Resumo:
Electrocardiogram (ECG) signals play a vital role in diagnosing heart conditions, but they are often corrupted by various types of noise, such as baseline wander, muscle artifacts, and electrode motion. These artifacts can obscure important diagnostic information and hinder the accurate analysis of ECG signals. This work presents a novel approach for ECG signal quality classification and noise filtering using the TiDE (Time-series Dense Encoder) architecture. A compact version, the TiDE Classifier, effectively identifies clean ECG signals, while the TiDE Filter removes noise from signals, by training on synthetic noisy data. The TiDE Filter outperforms traditional filtering methods, particularly in noisy conditions, improving the ECG analysis. This approach demonstrates the potential of deep learning for enhancing ECG signal quality, with promising applications in automated signal processing and medical diagnostics.
Banca examinadora:
- Prof. Dr. João Paulo do Vale Madeiro (MDCC/UFC – Orientador)
- Prof. Dr. César Lincoln Cavalcante Mattos (MDCC/UFC – Coorientador)
- Prof. Dr. Luís Otávio Rigo Júnior (UNILAB)