| Summary: | 碩士 === 國立成功大學 === 醫學工程研究所 === 83 === During the last decade echocardiography has experienced widespread acceptance in the evaluation of cardiac disease. Computer-aided analysis of the left ventricular (LV) wall motion provides quantitative parameters as objective diagnosis reference beside the visualization. In this study we establish a quantitative analysis system of two-dimensional echocardiography (2-D echo) images and present a method for extraction the endocardial border and apply quantitative analysis to a series of LV boundaries further.
An effective image processing algorithm for detection the cavity boundary in sequential 2-D echo is proposed. The main step involved in this algorithm are threshold estimating automatically, contrast stretching, boundary detection, and border smoothing. Good results have also been obtained in patient studies. Measurement of motion relies primarily on changes in endocardial contours from end-diastole to end-systole by the sample point connection method and radial method, and normalized for heart size. The final result of regional function is obtained after all the sample point was divided into six regions equally.
Of 5 patients studied, we conclude the sample point connection method is superior for the evaluation of having coronary artery disease with 2-D echo and has 73.7% accuracy of coronary artery judgement.
|
|---|
