基于边缘和区域信息相结合的变分 水平集图像分割方法

 针对GAC模型和C-V模型分别存在对弱边缘和灰度渐进图像分割效果不理想以及演化效率低等问题,提出了一种基于边缘和区域信息相结合的变分水平集图像分割方法.结合了图像边缘梯度信息和区域全局信息的能量函数作为模型的外部能量项,引入内部变形能量约束水平集函数来逼近符号距离函数,省去了重新初始化水平集函数的过程,并融入了物体形状先验知识的附加约束信息,提高了分割精度.实验结果表明,论文所用方法对分割噪声弱边缘图像和灰度渐进图像具有一定的有效性和可行性.

Varitional Level Set Image Segmentation Method Based on Boundary and Region Information

Being GAC model and C-V model has the unsatisfactory segmental results and inefficient curve evolution against weak boundary and intensity inhomogeneity images respectively, an unproved varitional level set image segmentation method is proposed. Our model consists of an external energy term that integrate the image information from both the gradient and region, and an internal energy term forces the level set function to be close a signed distance function.Therefore completely eliminates the need of re-initialization procedure.The method incorporate the additional constraint information based object's prior knowledge, which can improve the segmentation accuracy. Experimental results show that the method is effectiveness and feasibility on segmenting the noisy blurry boundary and intensity inhomogeneity images.