Matching And Tracking Of 2D Cardiac Ultrasound Image Based On Feature Points

Cardiovascular disease is a serious threat to human health risk of disease,ranking first among all causes of death,and by the medical imaging technology of heart disease of prevention,analysis,diagnosis,has important clinical significance.The geometry and motion of the heart are complex,and the detection method of the cardiac characteristic movement is insufficient..In recent years,the ultrasonic speckle tracking imaging technology can extract and match the feature of the myocardium,so as to track the movement of myocardial tissue..By observing and recording the related information of myocardial movement,a two-or three-dimensional model is constructed to simulate the movement of myocardial tissue..The combination of hospital and related diagnosis and diagnosis means to determine the movement trend of myocardial tissue characteristic points to analyze the correlation of cardiac muscle to provide information for the prevention and diagnosis of cardiovascular disease..Based on the excellent characteristics of SIFT,this paper uses SIFT algorithm to study the image matching and point tracking,and the conclusion is verified by experiments..In this paper,we first introduce the preprocessing of ultrasonic image,then introduce the concept of image feature,and extract the feature points by correlation algorithm..Is a very important feature extraction and image registration method based on SIFT algorithm.This paper will focus on the introduction SIFT algorithm,the specific operation is divided into five steps,each step to be described in detail,combined with after removing the noise of image operation SIFT algorithm,extract the feature points and show that every feature point descriptor.After extracting feature points for image matching based on feature point descriptor and setting the threshold value,on the platform of MATLAB analysis the correlation matching.Finally analysis cycle between myocardium contraction,record relevant information,further the feature points of the myocardial motion coherence,matching accuracy,and its motion analysis,provide a more accurate method for the analysis of myocardial motion.