知识卡片
矩阵形式看见频率混合
一维抽取一半时,频域映射可写为 (1/N)F₍N/2₎D₂Fᴴₙ,化为 [I,I]/2。输出每个频率是两段原频谱对应分量的平均,例如N=16时Y[0]=(X[0]+X[8])/2。这个视角把混叠写成显式的不可分离求和,而不只是图像变难看。
这张卡片看不懂?
这张卡片已通过人工审查,不再接受失败反馈。
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
Using matrices, we can write decimation as ℓ↓2 = D2ℓ. We use the Fourier matrices FN and to denote the DFT and its inverse (which is the complex conjugate of the DFT) for signals of length N. Using this notation, the relationship between the original spectrum ℒ [u] and the sampled version, ℒ↓2 [u], is:
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:摘录中关于矩阵形式看见频率混合的定义、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:Using matrices, we can write decimation as ℓ↓2 = D2ℓ. We use the Fourier matrices FN and to denote the DFT and its inverse (which is the complex conjugate of the DFT) for signals of length N. Using this notation, the relationship between the original spectrum ℒ [u] and the sampled version, ℒ↓2 [u], is: