融入形状先验的马尔科夫图像分割

Markov Image Into Shape Prior Segmentation

在马尔科夫(MRF)图像分割框架中融合形状先验约束,把图像分割问题作为最大后验(MAP)估计的一个马尔科夫随机场,在本质上,相当于最小化吉布斯能量函数.然后通过通量最大约束将形状先验信息合并到吉布斯能量函数,最后用图割技术最小化使吉布斯能量函数达到最优解,促使分割轮廓接近给定的形状模板.实验结果表明,算法效率得到了提高,分割效果得到了很大的改善.

In the Markov （MRF） fusion shape prior constraint framework for image segmentation, the image segmentation problem as a maximum a posteriori （MAP） estimation of a Markov random field, in essence, is equivalent to the minimization of the Gibbs energy function. Then the shape prior information is incorporated into the Gibbs energy function by flux maximum constraint, finally using graph cut techniques for minimizing the energy function to achieve Gibbs optimal solution, the segmentation contour close to the given shape template. The experimental results show that, the efficiency of the algorithm is improved, the segmentation effect is greatly improved.