【作者】 魏颖； 徐心和； 贾同； 赵大哲；

【Author】 WEI Ying1, XU Xin-he1, JIA Tong1,2, ZHAO Da-zhe1,2 (1.College of Information Science and Engineering, Northeastern University, Shenyang 110004, China; 2.The Software Center, Northeastern University, Shenyang 110179, China)

【机构】 东北大学信息科学与工程学院；

【摘要】 对Chan-Vese水平集图像分割方法进行分析和改进,提出了结合全局区域均值和局部边界信息的水平集改进算法,应用于肺部CT图像分割和肺结节检测。在图像分割的目标函数中,在Chan-Vese方法基础上,引入局部边界统计特性能量项,以利于提高肺部医学图像分割的准确率和分割速度。实验表明,该方法可以很好地分割出肺实质轮廓和肺结节病灶区域,在分割速度上比Chan-Vese方法有了明显的提高,检测结果不依赖于初始设置,将人工交互降至最低,有利于实现CT图像肺结节自动检测。更多还原

【Abstract】 The active contour method proposed by Chan-Vese was studied and improved. A new level set method was presented, which combined average values of global regions and local boundary information, and it was used for lung nodules detection for CT images. In the energy function, local boundary information was included as a fitting term in the segmentation model based on Chan-Vese method. It was effective for increasing the segmentation correct ratio and speed for lung CT image. The experiment results verify that it is a robust way for the segmentation of lung body and the detection for lung nodules, and the speed of calculation is faster than that of Chan-Vese’s. Moreover, the detection results are independent on the initial settings, so the manual mutuality can be reduced to the minimum, and it is an effective method for CT image lung nodule detection automatically.更多还原