摘要
目的改进外部力场以提高梯度矢量流(gradient vector flow,GVF)活动轮廓模型在超声斑点噪声环境下的边缘提取的稳定性和准确性。方法用角点判断函数和梯度控制函数调节平均曲率流(mean curvature vector flow,MCVF)扩散和各向异性扩散程度,生成新的外部力场,引导活动轮廓模型实现边缘提取。结果通过二值加噪图像和心脏超声图像的实验,验证了该方法的捕捉范围和处理噪声方面优于GVF。结论本文提出的边缘提取结果在超声医学图像方面鲁棒性强、初始化范围大、弱化噪声,同时能够保留边界以及尖角点,是适用于超声医学图像边缘提取的方法。
Objective To promote stability and accuracy of gradient vector flow （GVF） deformable models by improving external potential force field of ultrasound images under speckle noisy environments, Methods A new external force field was generated by using corner and gradient weigh function to control the mean curvature vector flow （MCVF） and anisotropic diffusion, then the deformable model was guided to achieve edge detection. Results According to the test on noise binary images and ultrasound images, it was proved that in this case the capture range was larger than that of GVF, and it was more insensitive to noise. Conclusion The results showed that the new method has the advantages of robustness, and large capture range. It reduces noise and at the same time preserves edges and sharp corner points in ultrasound images. It proves to be an effective method for edge detection of ultrasound images.
出处
《航天医学与医学工程》
CAS
CSCD
北大核心
2007年第3期205-208,共4页
Space Medicine & Medical Engineering
基金
四川省青年科学基金资助(04ZQ026-013)
关键词
超声图像
边缘提取
活动轮廓模型
各向异性扩散
平均曲率流
ultrasound images
edge detection
deformable model
anisotropic diffusion
mean curvature vector flow
