← 返回刚才阅读的位置

知识卡片

普通读书笔记卡 · 8896 · 生成模型:GPT-6

索贝尔组合平滑与求导

理解等级:中级

Sobel-Feldman 核把一轴的 [1,0,−1] 与另一轴的 [1,2,1] 组合,实现沿求导轴差分、沿垂直轴平滑。它是紧凑的可分离近似,目标包括改善方向响应。归一化和坐标符号取决于约定;它不保证所有频率都各向同性,也不保证所有噪声条件下最优。

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

      In two dimensions, we can use separable filters and build a partial derivative as (18.21) (18.22) This particular filter is called the Sobel-Feldman operator. The goal of this operator was to be compact and to be as isotropic as possible. The Sobel-Feldman operator can be implemented very efficiently as it can be written as the convolution with four small kernels, . Remember that b1 =

    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:摘录中关于索贝尔组合平滑与求导的定义、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:In two dimensions, we can use separable filters and build a partial derivative as (18.21) (18.22) This particular filter is called the Sobel-Feldman operator. The goal of this operator was to be compact and to be as isotropic as possible. The Sobel-Feldman operator can be implemented very efficiently as it can be written as the convolution with four small kernels, . Remember that b1 =