An FPGA Implementation of Multiclass Disease Detection From PPG

Research output: Journal Publications and ReviewsRGC 21 - Publication in refereed journalpeer-review

View graph of relations

Author(s)

Related Research Unit(s)

Detail(s)

Original languageEnglish
Article number6007604
Journal / PublicationIEEE Sensors Letters
Volume7
Issue number11
Online published5 Oct 2023
Publication statusPublished - Nov 2023

Abstract

Photoplethysmogram (PPG) signals react according to the volumetric change of blood, which is related to the total cardiovascular system. Hence, PPG signals have a very close relation with cardiovascular diseases (CVDs). As the number of CVD patients is increasing day by day, it is necessary to develop digital systems for early detection and treatment. In this letter, we have aimed to design a field programmable gate array (FPGA)-based hardware system that can detect various CVDs. We have chosen six different combinations of diseases from the PPG BP dataset. The selected signals are denoised using filters, and then features are extracted. Then, SVM classifier is used to determine the class of the preprocessed signals. For FPGA implementation, we have targeted Xilinx zynq ultrascale+. The system utilizes different logic and signal processing blocks to detect the result. The resource utilization analysis proves the efficacy of the selected FPGA board. The power utilization analysis depicts that the designed prototype requires 1.403 W of power, which is sensible due to the requirement of large number of logic blocks. The accuracy of the system of detecting multiclass diseases is 79.83%. In the future, the system can be further extended to detect other types of CVDs. © 2023 IEEE.

Research Area(s)

  • cardiovascular disease (CVD), field programmable gate array (FPGA), multiclass classification, photoplethysmogram (PPG), Sensor applications

Citation Format(s)

An FPGA Implementation of Multiclass Disease Detection From PPG. / Chowdhury, Aditta; Das, Diba; Hasan, Kamrul et al.
In: IEEE Sensors Letters, Vol. 7, No. 11, 6007604, 11.2023.

Research output: Journal Publications and ReviewsRGC 21 - Publication in refereed journalpeer-review