Noninvasive Blood Pressure Classification based on ECG with ResNet Algorithm - Details

author：

Liu, Wei (Liu, Wei.) ^[1] | Wang, Xin-Kang (Wang, Xin-Kang.) ^[2] | Wang, Liang-Hung (Wang, Liang-Hung.) ^[3]

Indexed by：

Abstract：

This　study　proposes　a　noninvasive　end-to-end　blood　pressure　(BP)　classification　model　based　on　electrocardiography　(ECG),　dividing　BP　into　three　categories,　namely,　normotension　(N),　prehypertension　(Pre),　and　hypertension　(H).　This　study　includes　three　parts:　(1)　preprocessing　ECG　signals,　including　cutting　signals　into　fixed-length　segments　and　denoising;　(2)　building　a　proper　ResNet　model　with　a　1D　convolutional　neural　network　(1D-CNN)　to　train　data;　and　(3)　evaluating　the　performance　of　the　classification　model.　The　classification　algorithm　can　identify　three　types　of　BP,　and　the　accuracy　rate　of　the　classification　algorithm　over　the　test　set　is　87.89%.　©　2021　IEEE.

Keyword：

Biomedical signal processing Blood Blood pressure Classification (of information) Convolutional neural networks Electrocardiography

Community：

[ 1 ] [Liu, Wei]Fuzhou University, College Of Physics And Information Engineering, Department Of Microelectronics, Fuzhou, China
[ 2 ] [Wang, Xin-Kang]Fujian Provincial Hospital, Fuzhou, China
[ 3 ] [Wang, Liang-Hung]Fuzhou University, College Of Physics And Information Engineering, Department Of Microelectronics, Fuzhou, China