Objective
Transform an image into the frequency domain and filter by zeroing frequency bands, then watch the inverse transform rebuild it.
The idea
Images have frequencies too: smooth areas are low, edges are high. Filter in the frequency domain and the inverse FFT brings it back to pixels.
Try this
- 1Take the FFT and locate the bright low-frequency centre.
- 2Zero the high frequencies and watch edges vanish.
- 3Remove the low frequencies and watch the detail-only image.
- 4Apply an inverse FFT and check the reconstruction.
Watch for
- Low frequencies dominate the image centre in the FFT.
- Edges and noise live in the high frequencies.
- Frequency filtering is multiplication in the FFT domain.
- The inverse FFT returns a spatial image.