Computer Vision Foundations · free preview

Phase, Amplitude and Transfer Functions

Phase carries the where and amplitude the strength, natural images share a 1/f amplitude law, and the transfer function classifies filters by which bands they pass.

1. Polar form splits a spectrum

Polar form splits a spectrum

2. The swap experiment

The swap experiment

3. Transfer functions classify filters

Transfer functions classify filters

4. Probing systems and eyes

Probing systems and eyes

5. What the swap experiment shows

A hybrid image is built from the amplitude spectrum of a mountain photo and the phase spectrum of a portrait, then inverted. What does the result look like?

6. Reading a transfer function

A filter's transfer function has gain near one at low frequencies, near zero at high frequencies, and a smooth fall in between. What does the filter do, and what is its DC gain close to?

7. Gain of an averaging kernel

A 3×33 \times 3 averaging kernel contains nine equal taps of 1/91/9. Write its DC gain ∣H[0,0]∣|H[0,0]|, the sum of all taps.

Practice this interactively

Free account · instant grading · spaced review that schedules itself.

Start this course for free