Intelligent estimation for the local pulse wave velocity based on the time-frequency image of pulse wave
SONG Wenxia
DENG Li
CHEN Zhi
WEI Jiayin
LU Youjun
ZHANG Jiajia
Abstract：To address the problem that the transit time(TT)method is susceptible to noise and reflected waves in the detection of carotid local pulse wave velocity(CLPWV),resulting in low accuracy of CLPWV estimation,we proposed an intelligent estimation method for the CLPWV estimation based on the time-frequency image of pulse wave.Firstly,a carotid local pulse wave propagation da-tabase(CLPWPD)in 4374 virtual adults with different ages and hemodynamic parameters was established.The central statistical mo-ments features and texture energy extraction features from the time-frequency image corresponding to the pulse waves were calculated.Then,each feature was combined separately with such parameters as age to get two comparison feature sets.Finally,different CLPWV prediction models were constructed based on five machine learning models.The results showed that the multi-layer perceptron(MLP)model performed the best overall,with the combination of energy feature set and MLP being less affected by noise and an average root mean squared error(RMSE)of 0.134.The research is expected to provide a new technical reference approach for the early and precise clinical screening and prevention of carotid artery atherosclerosis.
Keywords：Carotid arteryLocal pulse wave velocityTime-frequency image of pulse waveMachine learningDeep learning
Publication Date：2026-04-30
Online Publishing Date：2026-09-11(First online date of this platform, not the publication date of the document)
Pages：6( 157-162 )