局部熵驱动的生物医学图像分割偏移场恢复耦合模型

Coupled Model Driven by Local Entropy on Biological Medical Image Segmentation and Bias Recovery

为降低噪声影响同时恢复图像偏移场,提出一种基于局部熵信息的分割与偏移场恢复耦合模型.该模型在水平集理论的整体框架下将局部熵引入耦合模型,进而将其改造成全局凸函数,并利用Split-Bregman方法求得全局最优解.实验结果表明,文中模型可以准确、快速地分割灰度不均匀图像,同时可较好地恢复出图像的偏移场信息,对初始曲线和参数也具有较好的鲁棒性.

In order to reduce noise and recover bias field, we propose a coupled model for segmentation and bias recovery based on local entropy. Under the Level Set theory, this paper introduces the local entropy information into our coupled model, then rewrites the energy function as a global convex function solved by using the Split-Bregman method to obtain the global optimal solution. The experimental results show that our method can segment intensity inhomogeneous image accurately and rapidly, and recover the bias field of image greatly. Moreover, the model is robust to the initialization and parameters.