← 返回刚才阅读的位置

知识卡片

普通读书笔记卡 · 8907 · 生成模型:GPT-6

局部辛克需要明确窗口

理解等级:中级

用 sinc 的中心附近构造有限核,能降低理想重建的全局成本,却需明确截断窗口和归一化。书中对 Lanczos-1 的名称描述过简,不能把所有截取 sinc 中央瓣的实现视为同一个算法;比较插值器应记录实际核公式、支撑与边界方式。

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 nearest neighbor interpolation is the result of the convolution of ℓδ(t) with a box of width Ts. The linear interpolation, the interpolation kernel h(t) is a triangle of width 2Ts.8 Lanczos-1 interpolation consists in using as kernel only the central lobe of the sinc function. This provides an interpolation that is smooth using only a local neighborhood. Other interpolation methods such as cubic interpolation

    参考来源

    - 位置:《Foundations of Computer Vision (Antonio Torralba, Phillip Isola etc.)》第20章《图像采样与混叠》(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/c020b.xhtml) - 结论依据:摘录中关于局部辛克需要明确窗口的定义、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 nearest neighbor interpolation is the result of the convolution of ℓδ(t) with a box of width Ts. The linear interpolation, the interpolation kernel h(t) is a triangle of width 2Ts.8 Lanczos-1 interpolation consists in using as kernel only the central lobe of the sinc function. This provides an interpolation that is smooth using only a local neighborhood. Other interpolation methods such as cubic interpolation