不同滤波方法对超声图像分割的影响

采用各向异性滤波方法以及Gabor滤波方法对乳腺肿瘤超声图像进行处理，再使用snake方法以及levelset方法在设置相同参数的条件下，对过滤的图像分别进行分割。试验结果表明，不同的滤波方法在分害4目标图像的收敛度，边缘圆滑度及整体轮廓提取效果都有较大影响，为分割图像选取适当的滤波器提供了参考。

The Influences of Different Filtering Methods on the Segmentation of Ultrasonic Images

In order to explore the influences of different filtering methods on the segmentation of Ultrasonic Images, this thesis analyzes the Ultrasound Images of Breast tumor through the way of anisotropic filtering as well as the method of Gabor filtering. Then it uses the methods of snake and the level set to segment the filtered images respectively under the condition of the same parameters. The test results show that different filtering methods have great impacts on the Convergence of target image segmentation, the edge smooth and the overall effect of contour extraction, which has offered a reference to selection of the proper filtering.