Abstract: A segmentation method of B ultrasonic image based on the adaptive mask filtering and level set is presented，which can more accurately detect the treatment regions from the subtraction image of B ultrasonic image obtained before and after irradiating with high intensity focused ultrasound.By anatomizing and slicing up the fresh pork irradiated，the experimental results show that the actual injury is consistent with the result of the method.This method can provide help in monitoring clinical HIFU treatment，and has actual significance about enhancing the treatment efficacy of HIFU.

摘要： 提出了一种基于自适应掩模滤波和水平集方法相结合的B超图像分割方法，能较准确地从高强度聚焦超声（HIFU）辐照前后获得的B超图像的减影中检测出治疗区域。对HIFU辐照后的新鲜离体猪肉进行解剖切片，实际损伤情形基本与该方法所得结果相吻合。该方法可以帮助HIFU临床医生客观、准确地监控治疗过程，对提高HIFU疗效有实际意义。