The Pulse Displacement Detection Of The Carotid Artery Vascular Wall Based On The Improved Optical Flow Model

With the improvement of living standards,coronary heart disease,stroke,myocardial infarction and other diseases caused by the morbidity of cardiovascular and cerebrovascular disease has been increasing in recent years.Clinical studies have found that during the prophase of various kinds of cerebrovascular diseases,the arterial diameter will change significantly,which would directly result in the vascular wall elasticity weakened.Therefore,finding a suitable algorithm to detect the displacement of the vessel wall pulsation has become a viable research direction for the prevention of such diseases.However,the most commonly used Speckle Tracking Image algorithm is not accurate enough to detect the displacement of the carotid artery wall,and also the accuracy depends on the choice of plaque heavily.In this paper,an improved optical flow model is proposed,which is based on the classical Horn optical flow model and improved according to the characteristics of ultrasonic image motion.The classical optical flow estimation algorithm requires that the image brightness is constant and the displacement is small,but ultrasound images do not always meet these requirements.Therefore,this paper adds the L-K pyramid model hierarchical algorithm and the Hessian matrix to the discontinuous points in the classical optical flow algorithm to avoid the optical flow calculation error due to discontinuous tissue motion in the ultrasound image.After the improved algorithm is obtained,the algorithm is compared with the most commonly used speckle tracking algorithm in the current ultrasonic image motion detection,and the accuracy of the improved algorithm is analyzed.The displacement of the carotid artery image is affected by various disturbances,so the real displacement curve can not be obtained and directly used for the accurate verification of the algorithm.Therefore,this paper established a simulation model of carotid artery ultrasound,simulation of 30 groups of the beam is 16 lines,the scanning frequency of 1024Hz pulse wave along the vascular wall of the carotid artery B ultrasound sequence image.At the same time,the pulse motion of the ultrasonic image sequence is extracted by using the spot tracking,Horn light flow field model and the improved optical flow field model.The results are compared with the theoretical pulsation curve,and the error of several algorithms is quantitatively analyzed.The results show that the vibration curve of the pipe wall is closer to the theoretical curve,and the detection error is less than 10%.After the completion of the simulation verification,using the clinical carotid artery image to test wall displacement extraction effect of several algorithms.In summary,the carotid artery wall pulsation displacement curve extracted by the improved optical flow model is more accurate,which can provide a more accurate basis for the early diagnosis of cardiovascular and cerebrovascular disease.