Location:
Search - icm m
Search list
Description: This directory contains the MRF energy minimization software accompanying the paper
[1] A Comparative Study of Energy Minimization Methods for Markov Random Fields.
R. Szeliski, R. Zabih, D. Scharstein, O. Veksler, V. Kolmogorov,
A. Agarwala, M. Tappen, and C. Rother.
In Ninth European Conference on Computer Vision (ECCV 2006),
volume 2, pages 16-29, Graz, Austria, May 2006.
Optimization methods included:
1) ICM
2) Graph Cuts
3) Max-Product Belief Propagation
The TRW / TRW-S method is currently not included, but hopefully will be in a future version.
Instructions for compiling and using the software are included below.
CREDITS:
* MRF code, graph cut interface, and example code by Olga Veksler
* Graph cut library by Yuri Boykov and Vladimir Kolmogorov
* Belief propagation code by Marshall Tappen-This directory contains the MRF energy minimization software accompanying the paper
[1] A Comparative Study of Energy Minimization Methods for Markov Random Fields.
R. Szeliski, R. Zabih, D. Scharstein, O. Veksler, V. Kolmogorov,
A. Agarwala, M. Tappen, and C. Rother.
In Ninth European Conference on Computer Vision (ECCV 2006),
volume 2, pages 16-29, Graz, Austria, May 2006.
Optimization methods included:
1) ICM
2) Graph Cuts
3) Max-Product Belief Propagation
The TRW/TRW-S method is currently not included, but hopefully will be in a future version.
Instructions for compiling and using the software are included below.
CREDITS:
* MRF code, graph cut interface, and example code by Olga Veksler
* Graph cut library by Yuri Boykov and Vladimir Kolmogorov
* Belief propagation code by Marshall Tappen
Platform: |
Size: 47104 |
Author: fant |
Hits:
Description: 基于MRF图形的小波与分解 获取最粗尺度上的初始分割。使用EM算法必须有一个初值，因此我们首先使用K-均值聚类算法获取尺度J-1上的初始分割结果。
2.E步骤。使用MPL方法GMRF模型参数。
3.M步骤。使用估计出的参数，采用运算速度较快的迭代条件模式（ICM）通过最小化获取尺度上的优化的分割结果。
4.尺度内迭代。重复2和3知道满足某种准则，迭代停止。我们获得尺度n上的最终分割结果。
5.尺度间迭代。将尺度n的分割结果之间映射到最近的较细尺度n-1上，作为这个尺度的初始分割。重复4，直到获取最细尺度上的分割结果。(Wavelet and decomposition based on MRF graphics)
Platform: |
Size: 8698880 |
Author: Giovanna |
Hits: