FPGA-Based Multi Heart Diseases Classification System with the Aid of LabVIEW

Abstract

Electrocardiography (ECG) is one of the most important recording processes used in medicine; it provides a clear description of situation of the heart. The development of technological and computer science, which led to the emergence of high-resolution screens placed on the wrist and able to record the heart signal, increased the importance of developing a real time and portable multi heart diseases diagnosis system. In this paper, we propose an FPGA-based multi heart diseases classification system that identify eight different heart malfunctions depending on the standard ECG features. Our proposed ECG system, achieved with the aid of LabVIEW, consists of three parts: Acquisition System, Feature Extraction and Making Decision using two different classifiers; Threshold Decision (TD) and Numeral Virtual Generalizing Random-Access Memory (NVG-RAM) weightless neural network. The proposed classifiers were implemented using Verilog HDL and Xilinx Spartan 3AN FPGA. The FPGA mapping showed that TD classifier occupy 1% of the hardware platform slices with execution time of 16 ns, while the NVG-RAM classifier utilize 21% of the FPGA slices with an increase of the execution time equal to 12.81 mu s. On the other hand, the NVG-RAM outperforms the TD algorithm and the other proposed classifiers in the literature. In case of the experimental data, the probability of correct classification (PCC) of heart conditions was 100% for NVG-RAM and 98.84% for TD classifier. Whereas, the success rates in case of generated data for the executed TD and NVG-RAM classifiers were 98% and 100 %, respectively.