Forward fft:
$$ {\tt fft}(x+iy)=X + i Y = \sum(x +i y) (W_r + i W_i)\\=\sum (x W_r - y W_i ) + i (x W_i + y W_r)$$
Inverse fft:
$$ {\tt ifft}(X+iY)= \frac{1}{N}\sum(X +i Y) (W_r - i W_i)\\= \frac{1}{N} \sum (X W_r + Y W_i ) + i (-X W_i + Y W_r)\\ =\frac{1}{N}\sum \left(XW_r - (-Y) W_i\right) +(-i) \left(XW_i + (-Y)W_r\right) \\ = \frac{1}{N}\left[{\tt fft}(X-iY) \right]^* = \frac{1}{N}\left[{\tt fft}((X+iY)^*)\right]^*$$
또는
$$ \frac{1}{N} {\tt fft}(Y+iX)= \frac{1}{N} \sum(Y +i X) (W_r + i W_i)\\=\frac{1}{N} \sum (Y W_r - X W_i ) + i (Y W_i + X W_r)\\= y + i x$$
728x90
'Image Recognition > Fundamental' 카테고리의 다른 글
| FFT 구현 (0) | 2024.07.31 |
|---|---|
| CLAHE (2) (1) | 2024.06.26 |
| Approximate Distance Transform (0) | 2024.06.02 |
| Graph-based Segmentation (1) | 2024.05.26 |
| Linear Least Square Fitting: perpendicular offsets (0) | 2024.03.22 |
