基于水平集的闪光照相图像分割算法

针对Chan-Vese(CV)模型局部控制能力差的缺点,将基于区域的CV模型和分割曲线的局部信息结合起来,提出了一种新的水平集图像分割算法。该算法以CV法的分割曲线为初始曲线,以获得全局收敛性,在后继分割中引入分割曲线的局部信息,以提高模型对图像中微弱信号的分割能力。对闪光照相图像的数值实验表明,该算法噪声抵抗能力强,对初始轮廓位置不敏感,能实现对含细长拓扑结构和微小孔洞的弱边界闪光图像的自动分割。

Radiography image segmentation model based on level set

For it is difficult to deal with automatic segmentation of noisy weak edge radiography image with traditional Chan-Vese(C-V) algorithm, a new image segmentation algorithm for radiography based on the level set model has been explored. This algorithm uses the segmentation of C-V algorithm as the initial contours, and then constructs a speed function with both local information of active contour and global information of C-V method. It is proved that this model is robust to initial contours' positions and can segment noisy weak edge radiography image automatically. Besides, the testing for segmentation of the image with slender topological structure and tiny inner hole shows its significant effect.