| Name: | Description: | Size: | Format: | |
|---|---|---|---|---|
| 1.58 MB | Adobe PDF |
Advisor(s)
Abstract(s)
Este documento foi redigido no âmbito da Tese, do Mestrado em Engenharia Informática na área de Tecnologias do Conhecimento e Decisão, do Departamento de Engenharia Informática, do ISEP, cujo tema é classificação de sons cardíacos usando motifs.
Neste trabalho, apresenta-se um algoritmo de classificação de sons cardíacos, capaz de identificar patologias cardíacas. A classificação do som cardíaco é um trabalho desafiante dada a dificuldade em separar os sons ambiente (vozes, respiração, contacto do microfone com superfícies como pele ou tecidos) ou de ruído dos batimentos cardíacos.
Esta abordagem seguiu a metodologia de descoberta de padrões SAX (motifs) mais frequentes, em séries temporais relacionando-os com a ocorrência sistólica (S1) e a ocorrência diastólica (S2) do coração. A metodologia seguida mostrou-se eficaz a distinguir sons normais de sons correspondentes a patologia. Os resultados foram publicados na conferência internacional IDEAS’14 [Oliveira, 2014], em Julho deste ano.
Numa fase seguinte, desenvolveu-se uma aplicação móvel, capaz de captar os batimentos cardíacos, de os tratar e os classificar. A classificação dos sons é feita usando o método referido no parágrafo anterior. A aplicação móvel, depois de tratar os sons, envia-os para um servidor, onde o programa de classificação é executado, e recebe a resposta da classificação.
É também descrita a arquitetura aplicacional desenhada e as componentes que a constituem, as ferramentas e tecnologias utilizadas.
This document was prepared as part of the Thesis of the MSc in Computer Science in the area of Knowledge and Decision Technologies, Department of Computer Engineering, ISEP. The theme is classification of heart sounds. In this dissertation we present an algorithm for heart sounds classification, able to identify cardiac pathologies. The classification of the heart sound is a challenging work due to the difficulty in separating heartbeat sound from the ambient sounds (voices, breathing, microphone contact with surfaces like skin or textiles) or noise. In this approach we use the methodology of discovery of frequent SAX patterns (motifs) in time series, relating them with systolic (S1) and diastolic (S2) heart events. The methodology was effective to distinguish normal sounds from pathologic sounds. The results were published in international conference IDEAS'14 [Oliveira, 2014], in July. We have also developed a mobile application, able to capture, process and classify heart beats. The mobile application, captures and processes the sounds, sends them to a server where the classification program is running, and receives the classification result. We also described the application architecture, its components as well as the tools and technologies used.
This document was prepared as part of the Thesis of the MSc in Computer Science in the area of Knowledge and Decision Technologies, Department of Computer Engineering, ISEP. The theme is classification of heart sounds. In this dissertation we present an algorithm for heart sounds classification, able to identify cardiac pathologies. The classification of the heart sound is a challenging work due to the difficulty in separating heartbeat sound from the ambient sounds (voices, breathing, microphone contact with surfaces like skin or textiles) or noise. In this approach we use the methodology of discovery of frequent SAX patterns (motifs) in time series, relating them with systolic (S1) and diastolic (S2) heart events. The methodology was effective to distinguish normal sounds from pathologic sounds. The results were published in international conference IDEAS'14 [Oliveira, 2014], in July. We have also developed a mobile application, able to capture, process and classify heart beats. The mobile application, captures and processes the sounds, sends them to a server where the classification program is running, and receives the classification result. We also described the application architecture, its components as well as the tools and technologies used.
Description
Keywords
Sons cardíacos Patologia Coração Classificação Padrões Heart Sounds Pathologies Heart Classification Motifs