水平集方法能有效地解决曲线演化过程中的拓扑变化问题，基于全局特性的活动轮廓模型采用水平集方法给出的迭代解，能有效进行分割。用这种方法对CR胸片图像进行分割研究，并根据CR图像的特点，给出了只需一次演化的方法：直接针对原图像可以得到肺部大致区域；而针对CR的局部标准偏差图像可以得到肋骨的边缘图像。实验结果表明，该方法能很好地对CR图像进行分割。

Level set method could efficiently solve the problem of the topologic change involved in the evolvement of the curve.Using the loop solution presented by level set method, active contour model based on global traits can segment objects efficiently.The paper uses this method to research segmentation technique in Chest Radiography Image.Taking account of the characteristic of Chest Radiography Image,a method with only once evolvement is presented here.We could get coarse lung area from original image and could get rib edge from local standard deviation image of Chest Radiography.The results verify that it is a robust way for Chest Radiography Image segmentation.