高强度聚焦超声图像的去标记算法研究

Study on the De-labeling Algorithm of High Intensity Focused Ultrasound Image

高强度聚焦超声(HIFU)技术是一种治疗肿瘤的微创消融技术,手术中目标肿瘤通常是由操作医师人工分割获得的,这种方法的效率较低,但在使用自动分割方法时,往往面临HIFU图像本身质量较差,且其上有专家标记等问题。针对此问题,利用超声图像的噪声特点,提出一种HIFU图像的去标记算法,以达到对HIFU图像预处理的目的。实验表明,此算法可以有效的去除图像的标记信息,提高图像的质量。

High Intensity Focused Ultrasound as a minimally invasive treatment of tumor ablation technology, in HIFU operation the target tumor is usually determined by the doctors manual segmentation, and the efficiency of this method is not high, however, in the use of automatic segmentation method often face HIFU image quality is poor, and it has expert label and other issues. To overcome this problem, this paper proposed a de-labeling algorithm based on the noise feature of ultrasonic image, and the algorithm can achieve the purpose of preprocessing the HIFU image. The experiment results show that the algorithm.