基于改进自仿射映射系统与参数活动轮廓的医学图像分割算法  

Improved self-affine mapping system and parametric active contour based medical image segmentation algorithm

作　　者：赵立川[1] 高阳[2] 

机构地区：[1]川北医学院附属医院,四川南充637000 [2]南方医科大学生物医学工程学院,广州510515

出　　处：《计算机应用研究》2017年第2期616-620,共5页Application Research of Computers

基　　金：国家自然科学基金资助项目(61402080)

摘　　要：Snake模型分割图像时要求初始化轮廓线位于目标图像特征附近,且处理弱边界与深度凹陷区域的能力较弱,为此提出一种基于改进自仿射映射系统与参数活动轮廓的医学图像分割算法。首先,使用高斯滤波器对给定图像进行平滑处理并计算其小波系数;然后,在每个小波尺度的子矩阵中定义一些自仿射映射;之后,将不同小波尺度对应的子力叠加以获得自仿射力;最终,基于动态力公式引导Snake模型变形。基于医学图像的仿真实验结果表明,本算法对于医学图像的分割性能较好,有效地提高了Snake模型对弱边界与深度凹陷区域的处理能力。Snake model requires that the initial contour locates near the target image features while segmenting the image, and it is weak at processing the weak edge and deep concave region. Aimed at that shortcoming, this paper proposed an improved self-affine mapping system and parametric active contour based medical image segmentation algorithm. Firstly, it smoothed the given image by a Gaussian filter and computed itswavelet coefficients. Then, it defined some self-affine mappings in the sub- matrix of each wavelet scale. Next,it fused all sub-forces of each wavelet scale to final self-affine force. Lastly,it used dynamic force formulation to guide the evolution of the Snake model. Medical images based simulation experimented results show that the proposed algorithm performs better to medical image segmentation and improves the process capacity of the weak edge and deep concave region of Snake model.

关 键 词：医学图像 参数活动轮廓 高斯滤波器 小波系数 自仿射映射 

分 类 号：TP391.41[自动化与计算机技术—计算机应用技术]