摘要
由于受成像原理的限制,导致超声图像对比度低、边界模糊,因此基于边界的水平集分割效果很不理想。为了提高超声图像的分割精度和分割效率,提出了一种梯度信息与区域信息相结合的水平集分割算法。首先对基于边界的距离正则化水平集演化(DRLSE)模型进行改进,将区域信息引入到边界指示函数中,并用改进后的边界指示函数代替DRLSE模型中的边界指示函数,最后,得到一个梯度与区域信息相结合的水平集演化模型。结果表明,本文中的模型能准确分割甲状腺肿瘤超声图像,且在分割效率和分割精确度方面均比DRLSE模型有所提高。
Because of the restriction of imaging principle , ultrasound images led toare always with low contrast and weak boundaries , segmentation effect of level set based on edge was not ideal .In order to improve segmentation precision and efficiency of ultrasound images , new a novel level set segmentation algorithm was proposed combining gradient information with regional information was proposed .Firstly, distance regularized level set evolution ( DRLSE) model based on boundary was improved , regional information was put into boundary indicator function .And then, the improved boundary indicator function was used instead of DRLSE model ' s.Finally, a level set evolution model combining gradient information with regional information was obtained .The experimental results show that the model can accurately segment ultrasound images of thyroid tumor and the segmentation efficiency and precision are higher than DRLSE model .
出处
《激光技术》
CAS
CSCD
北大核心
2016年第1期126-130,共5页
Laser Technology
基金
河北大学医工交叉研究中心开放基金资助项目(BM201103)
关键词
图像处理
图像分割
距离正则化水平集演化模型
边界指示函数
梯度信息
区域信息
image processing
image segmentation
distance regularized level set evolution model
boundary indicator function
gradient information
region information