知识卡片

普通读书笔记卡 · 9159.a · 生成模型:GPT-6

高斯差分可近似尺度拉普拉斯

理解等级:高级

相近尺度高斯之差在适当缩放下近似尺度归一化LoG;常见约1.6比例是具体离散尺度选择,不是唯一理论必要常数。它们突出局部二阶结构,若要形成边缘仍需明确零交叉或其他选择规则。

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

      Differences of Gaussians can also be used to approximate the Laplacian of Gaussian (LoG), which is the sum of the second derivatives of a Gaussian kernel. Here, one kernel is roughly 1.6 times wider than the other. The band-pass characteristic of DoG and LoGs are important as they highlight high-frequency information such as edges yet suppress high-frequency noise in the image.

    参考来源

    - 位置:《Introduction to Autonomous Robots Mechanisms, Sensors, Actuators, and Algorithms (Nikolaus Correll, Bradley Hayes etc.)》第08章《视觉》(来源项目:a19c7933-4eda-4176-aebd-30fe7e1ad292;源文件:extracted/Introduction to Autonomous Robots Mechanisms, Sensors, Actuators, and Algorithms (Nikolaus Correll, Bradley Hayes etc.)/_epub-src/OEBPS/xhtml/chapter_8.xhtml) - 结论依据:摘录中关于高斯差分可近似尺度拉普拉斯的定义、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:Differences of Gaussians can also be used to approximate the Laplacian of Gaussian (LoG), which is the sum of the second derivatives of a Gaussian kernel. Here, one kernel is roughly 1.6 times wider than the other. The band-pass characteristic of DoG and LoGs are important as they highlight high-frequency information such as edges yet suppress high-frequency noise in the image.