A Tracking Method of Regional Myocardium on Ultrasonic RF Echo Signal Based on Active Net Model

The assessment of kinetic performance for a regional myocardium based on quantitative measures is important for disease diagnosis. Previously, we have discussed the use of ultrasonic radio frequency (RF) signals to improve tracking accuracy and proposed the hybrid automatic tracking method (HATM) for regional myocardial motion. However, the tracking accuracy of HATM is not sufficient in the region near endocardium. In this paper, we propose a new tracking method using Active Net model to improve tracking accuracy of HATM. Active Net is generally known as an object extraction method from an image using energy function optimization. Our method utilizes this model in ultrasonic RF echo signal space of M-mode. And, to use Active Net for tracking myocardial motion, it is need that suitable the default position of Active Net control points. To set the default positions and constraint positions, we employ the cardiac cycle and the depth positions of the endocardium in M-mode image, which are extracted by HATM. And we defined two energy functions as Active Net energy, which are the internal energy function and the potential energy function. The internal energy function is that maintains Active Net shape smoothly, and the potential energy function is that minimizes distance between the depth position of maximum value of correlation coefficient and position of Active Net control point. The shape of Active Net is changed by the energy minimization, and final Active Net shape is to be tracking result. Finally, the experimental results with clinical subjects show that the performance of the proposed method is empirically superior to the conventional method.