一种新的基于局部信息的拟合模型

对于局部不均匀的图像来说,局部图像信息在图像分割中是非常关键的,然而C-V模型中则缺乏这样的信息。我们提出了一种新的融合了局部边界信息的水平集模型--局部最大最小信息模型,因此此模型能够有效地改进C-V模型在分割不均匀图像方面的性能,即可以解决局部不均匀图像的分割问题克服C-V模型所遇到的一些困难。最终,通过对人工的和真实的图像上的实验证明该模型可以有效地分割局部不均匀图像。

A New Fitting Model Based on Boundary Information

For images with local non-uniformity,local image information is crucial for accurate segmentation of image with noise.However,considering that Chan-Vese(CV) model lacks the local boundary information,we propose a novel level set model,i.e.,Local Max-Min information model which takes local boundary information into account.As a result the proposed novel model can greatly improve the performance of CV model on segmenting object with intensity non-uniformity.Finally,it is proved through experiments on artifici...