← 返回刚才阅读的位置

知识卡片

普通读书笔记卡 · 8923.c · 生成模型:GPT-6

高斯金字塔递归减分辨率

理解等级:中级

令g₀为原图,每层先用归一化[1,4,6,4,1]/16沿两轴平滑,再抽取2倍,得到gₖ₊₁=Gₖgₖ。宽高各减半使二维像素数约减为1/4。不同层对应越来越宽的原图感受范围,但短二项式滤波是Gaussian平滑的近似,仍存在残余别名。

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

      We'd like to make a recursive algorithm for creating a multiresolution version of an image. A Gaussian filter is a natural one to use to blur out an image, since the multiple successive application of a Gaussian filter is equivalent to application of a single, wider Gaussian filter.

    参考来源

    - 位置:《Foundations of Computer Vision (Antonio Torralba, Phillip Isola etc.)》第23章《图像金字塔》(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/c023.xhtml) - 结论依据:摘录中关于高斯金字塔递归减分辨率的定义、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:We'd like to make a recursive algorithm for creating a multiresolution version of an image. A Gaussian filter is a natural one to use to blur out an image, since the multiple successive application of a Gaussian filter is equivalent to application of a single, wider Gaussian filter.