← 返回起始笔记位置

知识卡片

普通读书笔记卡 · 8878.b · 生成模型:GPT-6

相关与卷积相差核中心反转

理解等级:中级

实值图像本章约定下,相关用Σh[k,l]ℓ[n+k,m+l],卷积使用负偏移。相关适合把模板按原朝向与局部图块比较,卷积对冲激输出原核;中心对称核时两者一致。复值相关还需另行说明共轭约定。

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

      where the sum is done over the support of the filter h, which we assume is a square (–N, N) × (–N, N). In the convolution, the kernel is inverted left-right and up-down, while in the cross-correlation is not. Remember that the convolution between image ℓin and kernel h is written as follows:

    参考来源

    - 位置:《Foundations of Computer Vision (Antonio Torralba, Phillip Isola etc.)》第15章《线性图像滤波》(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/c015b.xhtml) - 结论依据:摘录中关于相关与卷积相差核中心反转的定义、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:where the sum is done over the support of the filter h, which we assume is a square (–N, N) × (–N, N). In the convolution, the kernel is inverted left-right and up-down, while in the cross-correlation is not. Remember that the convolution between image ℓin and kernel h is written as follows: