基于三阶样条曲线拟合的改进Snake模型在医学超声图像分割中的应用

鉴于医学超声图像所具有的复杂性,采用基于样条曲线拟合的改进Snake模型来实现对心脏超声图像的分割.为了克服传统Snake模型对初始轮廓的依赖性,采用扇形法获得靠近目标边界的初始轮廓点,并且得到的轮廓点是有序点集.能量最小化过程运用贪婪算法来获得图像的特征边缘点,最后采用三阶样条拟合的方法来获得连续的图像边缘.实验结果表明,采用本方法可以获得连续、封闭的边缘曲线,能够较好地将目标从图像中提取出来.

Improved Snake Model Based on Spline Curve Fitting for Medical Ultrasound Image Segmentation

As medical ultrasound image segmentation’s complexity,this paper adopts improved Snake model based on spline curve fitting for cardiac ultrasound image segmentation.In order to decreasing the Snake ’dependence on initial contour,this paper adopts fan-shaped method for the initial contour point selection.By using this method,the initial contour points near the object boundary can be obtained,and these points are sequential.Greedy algorithm is used to obtain the characteristic edge points of the image in the process of energy minimizing.On the basis of that,a three order spline curve fitting method is adopted to get continuous edge of the image.The result shows that to use the method presented in this paper can get a continuous and close edge,thus the target object can be well extracted from the ultrasound image.