← 返回起始笔记位置

知识卡片

普通读书笔记卡 · 8911.c · 生成模型:GPT-6

锐利频带可能造成光晕

理解等级:中级

理想低通的振荡核在突变边缘附近可产生振铃;窗函数降低一部分波纹后,图中仍可见光晕。正系数的二项式平滑能避免这类负瓣振铃,却更模糊且抗混叠不完全。应同时评价别名、锐度和边缘伪结构,不只看截止曲线。

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

      The images in figure 21.15 show the output of three different low-pass anti-aliasing filters. Note how the ideal low-pass filter introduces ringing artifacts in the output image (figure 21.15[a]). Despite that it removes frequencies that will produce aliasing after downsampling, it does that at a high cost because it introduces undesired image structures, which is exactly what we want to prevent with the anti-aliasing filter.

    参考来源

    - 位置:《Foundations of Computer Vision (Antonio Torralba, Phillip Isola etc.)》第21章《图像下采样与上采样》(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/c021.xhtml) - 结论依据:摘录中关于锐利频带可能造成光晕的定义、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:The images in figure 21.15 show the output of three different low-pass anti-aliasing filters. Note how the ideal low-pass filter introduces ringing artifacts in the output image (figure 21.15[a]). Despite that it removes frequencies that will produce aliasing after downsampling, it does that at a high cost because it introduces undesired image structures, which is exactly what we want to prevent with the anti-aliasing filter.