FPGA Implementation of PPG-Based Cardiovascular Diseases and Diabetes Classification Algorithm

Photoplethysmogram is a noninvasive technique used to detect volumetric changes in the blood. Cardiovascular diseases, related to heart and blood supply problems, are one of the largest causes of death in the world. Our study explored the possibility of classifying different cardiovascular diseases using photoplethysmogram signals for quick diagnosis. Using the support vector machine technique, the classification is done at the software level, while Xilinx Zynq 7000 field-programmable gate array (FPGA) chip is utilized for hardware design. The overall accuracy for detecting cerebral infarction and cerebrovascular disease is 93.48% and 96.43%, respectively, using eleven features. In addition, diabetes which is linked to cardiovascular diseases is classified, and an accuracy of 88.46% is achieved. Considering the PPG signal with a higher signal quality index, the overall accuracy for all the diseases can be further increased. The resource and power utilization of the implemented system is analyzed, which shows that 0.693 W power is required. The developed prototype can be further extended as a point-of-care system for cardiovascular disease detection. © King Fahd University of Petroleum & Minerals 2024.