知识卡片

普通读书笔记卡 · 8878 · 生成模型:GPT-6

循环卷积把索引按周期回绕

理解等级:中级

长度N循环卷积用模N索引将越界部分回绕,结果可看作周期N的序列;其矩阵为循环结构。它不同于零扩展的完整线性卷积,尾部可能叠回开头。选择周期模型是边界假设,不代表原始照片两侧真实相接。

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

      convolving a finite length signal with a kernel, we have seen that it can be convenient to extend the signal beyond its boundary to reduce boundary artifacts. This trick is also very useful analytically and conduces to a special case of the convolution: the circular convolution uses circular padding. The circular convolution of two signals, h and ℓin, of length N is defined as follows:

    参考来源

    - 位置:《Foundations of Computer Vision (Antonio Torralba, Phillip Isola etc.)》第15章《线性图像滤波》(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/c015b.xhtml) - 结论依据:摘录中关于循环卷积把索引按周期回绕的定义、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:convolving a finite length signal with a kernel, we have seen that it can be convenient to extend the signal beyond its boundary to reduce boundary artifacts. This trick is also very useful analytically and conduces to a special case of the convolution: the circular convolution uses circular padding. The circular convolution of two signals, h and ℓin, of length N is defined as follows: