Snake模型在乳腺肿瘤超声图像处理中的运用

目的对乳腺超声图像中的肿瘤进行边缘提取。方法鉴于医学超声图像的信噪比较低,用经典的边缘提取算法无法得到较好的结果,因此,Snake模型作为一种基于高层信息的有效目标轮廓提取算法而引起广泛的关注。在原始的Snake模型的基础上,本文针对超声图像的特点对它进行了一些改进。结果从上海第六人民医院采集到乳腺超声图像15幅。在进行了灰度分割、形态滤波等一系列预处理后,将改进后的Snake模型运用到边缘提取中来,并在这15幅图像中得到了比较好的分割结果。结论改进后的Snake模型可以将乳腺超声图像中肿瘤的边缘较好地提取出来,为乳腺肿瘤计算机辅助诊断提供了重要依据。

Application of Snake model for edge extraction breast tumor in ultrasound images

Objective To extract the edge of breast tumors in ultrasound images. Methods Because of the low signal-to-noise ratio (SNR) of medical ultrasound images, it was hard to get optimal result using classical edge extraction algorithms. As an effective edge extraction algorithm based on advanced information, Snake model attracts more and more attention. In this paper, Snake model was improved with the prior knowledge of ultrasound images from the original model. Results Fifteen ultrasound images of the breast tumor were collected from Shanghai Sixth People's Hospital. After a lot of pre-processing like gray-level segmentation and shape filtering,the improved Snake model was applied to the ultrasound images of breast tumor. This new method could extract the edge of the tumor in most test images and get better result than some classical methods. Conclusions The improved Snake model can extract the edge of breast tumors in ultrasound images and play an important part in computer-aided diagnosis algorithm of breast tumors.