B超图像HIFU损伤区域分割方法的研究

针对目前高强度聚焦超声（High-Intensity Focused Ultrasound，HIFU）治疗过程中图像分割环节耗时过长且依赖于临床医生对损伤区域的准确判断的问题。提出了一种新的自动分割方法，该方法面向HIFU治疗这一具体应用，由治疗前治疗计划拟定的靶点和靶区大小对治疗后B超图像划定感兴趣区域（ROI）并对区域内图像进行最大类间方差法（Otsu）阈值分析，然后根据分析得到的阈值信息对治疗后图像进行阈值水平集分割。实验结果证明，与基于区域阂值的分割方法进行比较，该方法的敏感性和准确度分别可达98．67％和93．52％。

Segmentation Method of HIFU Lesion Areaon Ultrasound B-modeImage

Focus on the fact that it takes too long in the step of image segmentation during HIFU treatment and the segmentation relies on manual identification of the lesion region. A new automatic segmentation method is proposed for HIFU treatment. This method delimits ROI region on the ultrasound B-mode image after treatment based on the target spot and target region planned before treatment, Otsu threshold analysis of the ROI region and then segment the B-mode ultrasound image after treatment based on threshold level set method according to the information of threshold. The experimental results show that compared with the segmentation method based on areathreshold, the sensitivity and accuracy of this method can reach 98.67% and 93.52%,respectively.