基于遗传算法的水平集超声图像分割

鉴于医学超声图像的信噪比较低,用经典的边缘提取算法无法得到较好的结果,因此,提出超声图像自动分割的一种新方法.其基本思想是在水平集分割方法的基础上,以能量函数作为评价函数,把图像分割问题变成一个优化问题,利用遗传算法的寻优高效性,搜索到能使分割质量到达最优的轮廓曲线.应用此方法对肝脏超声图像进行肿瘤的边缘提取,得到比较好的结果,从而完成图像的自动分割.

Level Set Method in the Ultrasound Image Segmentation Based on Genetic Algorithm

Where as the signal to noise ratio of medical ultrasound images is quite cow, it can not get optimal result using classical edge extraction algorithm. So, the article presents an algorithm for automatic segmentation of ultrasound images. The segmentation problem is formulated as one of optimization, and the energy function is taken as the segmentation quality criteria, which also fits well with GA. The GA efficiently searches the contours in order to obtain the optimal thresholds. And this new method can extract the edge of the lesion and get preferable result, and finish the automatic segmentation.