【摘要】 为超声引导的HIFU治疗系统设计一种自动检测凝固性坏死损伤区域的大小形状及位置的辅助诊断方法,本文提出一种基于水平集的HIFU损伤区域自动探测方法,利用超声剪影图像所含灰度变化信息,自动识别热损伤发生的位置,设置水平集函数的初始值,进而对热损伤区域的的边界进行检测.该算法分割出的组织损伤区域形状与大小均与组织剖面所见损伤区域相符,对连通或非连通区域均可进行很好的分割.更多还原

【Abstract】 In order to design a computer-aided approach for HIFU treatment systems to monitor HIFU lesion’s size,shape and location automatically,the paper proposes a level-set based detection method to detect the edge of HIFU thermal lesion area.This method uses grey-level change information in the subtraction image to locate the position of thermal lesion and initialize level set function automatically.Results of experiments show that the shape and size of the segmentation area all agree with that on the section plane of the tissue.It has been proved that this method can detect the thermal lesion area with weak edge correctly and automatically no matter the area is connected or non-connected.更多还原