← 返回刚才阅读的位置

知识卡片

普通读书笔记卡 · 8897.a · 生成模型:GPT-6

零交叉轮廓也有边界条件

理解等级:高级

LoG 零交叉曾用于边缘定位,但原书指出它作为通用边缘检测并不可靠。关于“零交叉形成闭合轮廓”的直觉还需限定在规则、连续的零水平集及其定义域条件;图像边界、临界点和离散阈值会产生开放段或分叉,不能据此强制补成闭合物体。

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

      Edges can be located as the zero-crossings in the Laplacian output. However, this way of detecting edges is not very reliable. Zero crossings of an image form closed contours. Figure 18.19 compares the output of the first order derivative (figure 18.19[b]) and the Laplacian (figure 18.19[d]) on a simple 1D signal (figure 18.19[a]). The local maximum of the derivative output (figure 18.19[c]) and the zero

    参考来源

    - 位置:《Foundations of Computer Vision (Antonio Torralba, Phillip Isola etc.)》第18章《图像导数》(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/c018.xhtml) - 结论依据:摘录中关于零交叉轮廓也有边界条件的定义、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:Edges can be located as the zero-crossings in the Laplacian output. However, this way of detecting edges is not very reliable. Zero crossings of an image form closed contours. Figure 18.19 compares the output of the first order derivative (figure 18.19[b]) and the Laplacian (figure 18.19[d]) on a simple 1D signal (figure 18.19[a]). The local maximum of the derivative output (figure 18.19[c]) and the zero