基于Chan-Vese水平集的梯度加速分割模型

Advanced gradient segmentation model based on Chan-Vese level set method

为了提高图像分割的速度和精度,提出了一种新的基于Chan-Vese水平集模型(C-V模型)的梯度加速分割模型。首先,在C-V模型的能量函数中加入一个内部能量项,抵消演化过程中水平集函数和符号距离函数的偏差,从而消除分割中周期性重新初始化的过程;其次,提出了梯度加速项,通过感兴趣区域的图像特征,快速得到该区域的边界,且能够提高弱边界的分割精度。实验证明,提出的方法不仅能够加速特定区域的分割、提高分割精度,还能保持分割过程的稳定性。

In order to increase the speed and precision of image segmentation, proposed a novel gradient advanced segmentation model based on Chan-Vese level set method. First, added an internal energy term in order to counteract the discrepancy of the level set function and the signed distance function during iteration, so eliminated the troublesome re-initialization process. Second, proposed the gradient-advanced term, which could attain the desired boundaries faster than others if there was more than one object in the detection area. Moreover, the gradient-advanced term could also increase the segmentation precision of weak boundaries. A number of experiments prove the validity and robustness of the model in segmenting exact areas and maintaining the stability of evolvement.