Are spatial and global constraints really necessary for segmentation?
Are spatial and global constraints really necessary for segmentation?
复制标题
DOI：
10.1109/iccv.2011.6126219
复制
发表时间：
2011-11
期刊：
影响因子：
--
通讯作者：
Aurélien Lucchi;Yunpeng Li;X. Boix;Kevin Smith;P. Fua
中科院分区：
文献类型：
--
作者：
Aurélien Lucchi;Yunpeng Li;X. Boix;Kevin Smith;P. Fua
许多最先进的分割算法依赖于马尔可夫或条件随机场模型，旨在加强空间和全局一致性约束。这通常是通过向模型引入额外的潜在变量来实现的，这会大大增加模型的复杂性。因此，估计模型参数或计算最佳最大后验（MAP）分配成为一项计算成本很高的任务。在PASCAL和MSRC数据集上的一系列实验中，我们无法找到由于引入此类约束而显着提高性能的证据。相反，我们发现可以使用一个更简单的设计来实现类似的性能水平，这种设计基本上忽略了这些约束。这种更简单的方法利用相同的局部和全局特征来利用图像中的证据，但直接影响单个像素的偏好。虽然我们的调查并没有证明空间和一致性约束在原则上是无用的，但它指出了一个结论，即它们应该在更大的背景下得到验证。
Many state-of-the-art segmentation algorithms rely on Markov or Conditional Random Field models designed to enforce spatial and global consistency constraints. This is often accomplished by introducing additional latent variables to the model, which can greatly increase its complexity. As a result, estimating the model parameters or computing the best maximum a posteriori (MAP) assignment becomes a computationally expensive task. In a series of experiments on the PASCAL and the MSRC datasets, we were unable to find evidence of a significant performance increase attributed to the introduction of such constraints. On the contrary, we found that similar levels of performance can be achieved using a much simpler design that essentially ignores these constraints. This more simple approach makes use of the same local and global features to leverage evidence from the image, but instead directly biases the preferences of individual pixels. While our investigation does not prove that spatial and consistency constraints are not useful in principle, it points to the conclusion that they should be validated in a larger context.