一种针对灰度非均匀图像的混合分割框架

该文将基于区域的局部和全局活动轮廓技术有机结合起来,给出了一种针对灰度非均匀图像的分割框架。考虑到灰度非均匀特性对图像非轮廓象素点影响较大而对图像轮廓象素点影响较小这一事实,在该框架中引入了图像边缘停止函数来平衡局部和全局灰度信息之间的相互影响。基于该框架,给出了一个具体的实例模型,运用快速的模糊对偶算法获取实例模型的全局最优解。用一组图像进行实验,并通过对实验结果的定量分析,验证了该模型的合理性。

A Hybrid Segmentation Framework for Inhomogeneity Image

Based on local and global region-based active contours , an adaptive image segmentation framework with the capability of handling the intensity inhomogeneity is presented .Considering the fact that the intensity inhomogeneity mainly influences the non-contour pixels in the image and less for the pixels belonging to contour , a stop function of the image is applied to balance the influence of the local intensity information and the global intensity information in this framework .Based on this framework , an instancing model is gave , which is solved by the fuzzy dual method to obtain the globally optimal solution .Some image segment experiment results are presented and the quantificationally evaluate of the results proved that the framework is available .