RESUMEN
Abstract: Myocardial infarction is one of the leading causes of death worldwide. As it is life threatening, it requires an immediate and precise treatment. Due to this, a growing number of research and innovations in the field of biomedical signal processing is in high demand. This paper proposes the association of Reconstructed Phase Space and Artificial Neural Networks for Vectorcardiography Myocardial Infarction Recognition. The algorithm promotes better results for the box size 10 x 10 and the combination of four parameters: box counting (Vx), box counting (Vz), self-similarity method (Vx) and self-similarity method (Vy) with sensitivity=92%, specificity=96% and accuracy=94%. The topographic diagnosis presented different performances for different types of infarctions with better results for anterior wall infarctions and less accurate results for inferior infarctions.
Asunto(s)
Vectorcardiografía , Algoritmos , Redes Neurales de la Computación , Infarto del MiocardioRESUMEN
Myocardial infarction is one of the leading causes of death worldwide. As it is life threatening, it requires an immediate and precise treatment. Due to this, a growing number of research and innovations in the field of biomedical signal processing is in high demand. This paper proposes the association of Reconstructed Phase Space and Artificial Neural Networks for Vectorcardiography Myocardial Infarction Recognition. The algorithm promotes better results for the box size 10â¯×â¯10 and the combination of four parameters: box counting (Vx), box counting (Vz), self-similarity method (Vx) and self-similarity method (Vy) with sensitivityâ¯=â¯92%, specificityâ¯=â¯96% and accuracyâ¯=â¯94%. The topographic diagnosis presented different performances for different types of infarctions with better results for anterior wall infarctions and less accurate results for inferior infarctions.
