医学超声图像交互型目标检测  

Interactive Model of Detecting Object in Ultrasound Image

作　　者：王峥[1] 杨新[1] 施鹏飞[1] 

机构地区：[1]上海交通大学图像处理与模式识别研究所,上海200030

出　　处：《仪器仪表学报》2002年第z2期675-677,共3页Chinese Journal of Scientific Instrument

基　　金：国家自然科学基金资助项目(No.60072026)。

摘　　要：医学图像中要分割的对象常比较复杂,通常的图像自动检测方法很难得到理想的结果。在曲线演化图像分割方法的基础上,提出了一种基于水平集方法的人机交互模型。该模型不仅继承了水平集方法的拓扑变化的自适应性,而且还有良好的人机交互性能。在分割过程中,医生只要在图像的适当位置上加入少许几个标记点,就可以在医生的指导下对复杂的对象进行准确的检测。实验表明,交互模型具有良好的实用性。In the medical image processing,the desired detection object is often very complicated. It is difficult to acquire a satisfied result using a fully automatic image segmentation method. The segmentation guided by experienced doctors is the only possible solution to have reasonable results. We present an interactive model based on the level set method,and combine it with geodesic active region segmentation method. The model inherits the a-bility of topology adaptability of the level set method. Furthermore,doctors are only needed to put a few mark points on the suitable image positions,and then able to monitor the detection results. Our experiments show that the model is not only practical,but also reliable.

关 键 词：图像分割 人机交互模型 曲线演化 水平集方法 

分 类 号：TP2[自动化与计算机技术—检测技术与自动化装置]