← 返回刚才阅读的位置

知识卡片

普通读书笔记卡 · 8962.b · 生成模型:GPT-6

网格图描述空间邻接

理解等级:中级

图像MRF可把像素标签作为隐藏节点,用局部观测势表示每个标签与图像的相容性,用邻接势鼓励某种空间一致性。网格只是常用结构,MRF不只限二维像素;平滑标签也应允许真实边界,而不是强制所有相邻位置相同。

这张卡片看不懂?

已有反馈 · 0 条

    还没有人提交反馈。

    在整书笔记中的位置

    原书出处 · 1 处

    • · 查看原书依据

      To capture relationships over space, a two-dimensional (2D) structure is useful, called a Markov random field (MRF) [53]: Figure 29.9: Two-dimensional Markov random field. Then the joint probability over all the variables factorizes into this product: (29.4) where the first product is over all spatial neighbors i and j, and the second product is over all nodes i. An example of how Markov random field

    参考来源

    - 位置:《Foundations of Computer Vision (Antonio Torralba, Phillip Isola etc.)》第29章《概率图模型》(来源项目:a19c7933-4eda-4176-aebd-30fe7e1ad292;源文件:extracted/Foundations of Computer Vision (Antonio Torralba, Phillip Isola etc.)/_epub-src/OPS/xhtml/c029.xhtml) - 结论依据:摘录中关于网格图描述空间邻接的定义、比较或条件支撑卡片;原书表述不严谨的范围已在覆盖记录逐项限定,不把案例结果外推为普遍结论。 - 原始内容:To capture relationships over space, a two-dimensional (2D) structure is useful, called a Markov random field (MRF) [53]: Figure 29.9: Two-dimensional Markov random field. Then the joint probability over all the variables factorizes into this product: (29.4) where the first product is over all spatial neighbors i and j, and the second product is over all nodes i. An example of how Markov random field