FPGA-based Cloudification of ECG Signal Diagnosis Acceleration

Abstract

Recently, studies to analyze heart disease using ECG signals are emerging. The proposed platform generates multiple reference signals trained for individuals in real time by reducing the learning time. The data in the cluster is compressed by linear approximation to speed up diagnosis and reduce memory usage, allowing more diagnosis to be performed with limited resources. Platforms using FPGA can accelerate ECG signal diagnosis by adding hardware. As a result of diagnosing ECG signals of 10 people using the processor and accelerator, the execution time when using the accelerator was 71% lower than that when using the processor.