中文摘要：

医学图像分割是临床诊断的关键环节,分割结果将直接关系到后续对病灶的识别。C-V模型（Chan-Vese）大量应用于各种医学图像分割过程。围绕肝脏超声图像,针对传统C-V模型依赖初始轮廓及运算复杂耗时的特点,融合随机森林方法,提出一种基于边缘引导能量函数和局部约束特征的分割方法,利用随机森林节点生长和分类速度快的优势,在粗分割的基础上形成无需初始化的C-V模型,而后借助分类特征得到精准的肝脏区域及病灶分割结果。实验证明,经过优化的改进方法是可行有效的,对于图像中的组织和病灶区域能有效分割和提取，

英文摘要：

Medical image segmentation plays a key role in clinical diagnosis .The result will be directly related to the recognition of focus. C-V Model is widely used in medical image segmentation. Revolving around liver ultra sound image, a method is proposed based on random forests that we use edge energy function and local restricted feature to avoid re-initialization of the initial contour position. The classification feature can also be used to realize the fine segmentation. The result shows that the algorithm is effective and simple. The tissues and focus can be effectively extracted.