CS 4495 Computer Vision – A.

Bobick Frequency and Fourier Transform

EEE415 Digital Image Processing

Frequency domain filtering

Slides from Aaron Bobick and

Frank (Qingzhong) Liu
CS 4495 Computer Vision – A. Bobick Frequency and Fourier Transform

2-D Impulse and Sifting Property: Continuous

CS 4495 Computer Vision – A. Bobick Frequency and Fourier Transform

2-D Impulse and Sifting Property: Discrete

CS 4495 Computer Vision – A. Bobick Frequency and Fourier Transform

2-D Fourier Transform: Continuous

CS 4495 Computer Vision – A. Bobick Frequency and Fourier Transform

2-D Fourier Transform: Continuous

CS 4495 Computer Vision – A. Bobick Frequency and Fourier Transform

2-D Sampling and 2-D Sampling Theorem

CS 4495 Computer Vision – A. Bobick Frequency and Fourier Transform

2-D Sampling and 2-D Sampling Theorem

CS 4495 Computer Vision – A. Bobick Frequency and Fourier Transform

2-D Sampling and 2-D Sampling Theorem

CS 4495 Computer Vision – A. Bobick Frequency and Fourier Transform

Aliasing in images
CS 4495 Computer Vision – A. Bobick Frequency and Fourier Transform

Aliasing in images
CS 4495 Computer Vision – A. Bobick Frequency and Fourier Transform

Aliasing in Images: Example

In an image system, the

number of samples is
fixed at 96x96 pixels. If
we use this system to
digitize checkerboard
patterns …

Under-sampling
CS 4495 Computer Vision – A. Bobick Frequency and Fourier Transform

Image half-sizing
This image is too big to
fit on the screen. How
can we reduce it?

How to generate a half-

sized version?

S. Seitz
CS 4495 Computer Vision – A. Bobick Frequency and Fourier Transform

Image sub-sampling

1/8
1/4

Throw away every other row and

column to create a 1/2 size image
- called image sub-sampling
S. Seitz
CS 4495 Computer Vision – A. Bobick Frequency and Fourier Transform

Image sub-sampling

1/2 1/4 (2x zoom) 1/8 (4x zoom)

Aliasing! What do we do?

S. Seitz
 CS 4495 Computer Vision – A. Bobick Frequency and Fourier Transform

Gaussian (lowpass) pre-filtering

G 1/8
G 1/4

Gaussian 1/2
Solution: filter the image, then subsample
S. Seitz
CS 4495 Computer Vision – A. Bobick Frequency and Fourier Transform

Subsampling with Gaussian pre-filtering

Gaussian 1/2 G 1/4 G 1/8

S. Seitz
CS 4495 Computer Vision – A. Bobick Frequency and Fourier Transform

Compare with...

1/2 1/4 (2x zoom) 1/8 (4x zoom)

S. Seitz
CS 4495 Computer Vision – A. Bobick Frequency and Fourier Transform

Aliasing in Images: Example

Re-sampling
CS 4495 Computer Vision – A. Bobick Frequency and Fourier Transform

Aliasing in Images: Example

Re-sampling
CS 4495 Computer Vision – A. Bobick Frequency and Fourier Transform

2-D Discrete Fourier Transform and Its

Inverse

IDFT:

for x = 0, 1, 2…M-1 and y = 0, 1, 2…N-1

CS 4495 Computer Vision – A. Bobick Frequency and Fourier Transform

Properties of the 2-D DFT

relationships between spatial and frequency intervals
CS 4495 Computer Vision – A. Bobick Frequency and Fourier Transform

Properties of the 2-D DFT

translation and rotation
CS 4495 Computer Vision – A. Bobick Frequency and Fourier Transform

Properties of the 2-D DFT

periodicity
CS 4495 Computer Vision – A. Bobick Frequency and Fourier Transform

Properties of the 2-D DFT

periodicity
CS 4495 Computer Vision – A. Bobick Frequency and Fourier Transform

Properties of the 2-D DFT

Symmetry
CS 4495 Computer Vision – A. Bobick Frequency and Fourier Transform

Properties of the 2-D DFT

Fourier Spectrum and Phase Angle
CS 4495 Computer Vision – A. Bobick Frequency and Fourier Transform
CS 4495 Computer Vision – A. Bobick Frequency and Fourier Transform
CS 4495 Computer Vision – A. Bobick Frequency and Fourier Transform

Examples

Magnitudes are shown

CS 4495 Computer Vision – A. Bobick Frequency and Fourier Transform

Phase and Magnitude

•Fourier transform of a real

function is complex
• difficult to visualize
• instead, think of phase and
magnitude
•Phase = phase of the
complex transform
•Magnitude = magnitude of
complex transform
CS 4495 Computer Vision – A. Bobick Frequency and Fourier Transform
CS 4495 Computer Vision – A. Bobick Frequency and Fourier Transform

Example: Phase Angles

CS 4495 Computer Vision – A. Bobick Frequency and Fourier Transform

Example: Phase Angles and The Reconstructed

CS 4495 Computer Vision – A. Bobick Frequency and Fourier Transform
 CS 4495 Computer Vision – A. Bobick Frequency and Fourier Transform

This is the
magnitude
of cheetah
 CS 4495 Computer Vision – A. Bobick Frequency and Fourier Transform

This is the
phase of
cheetah
CS 4495 Computer Vision – A. Bobick Frequency and Fourier Transform
 CS 4495 Computer Vision – A. Bobick Frequency and Fourier Transform

This is the
magnitude
of zebra
 CS 4495 Computer Vision – A. Bobick Frequency and Fourier Transform

This is the
phase of
zebra
 CS 4495 Computer Vision – A. Bobick Frequency and Fourier Transform

Reconstruct
with zebra
phase,
cheetah
magnitude
 CS 4495 Computer Vision – A. Bobick Frequency and Fourier Transform

Reconstruct
with
cheetah
phase, zebra
magnitude
CS 4495 Computer Vision – A. Bobick Frequency and Fourier Transform

2-D Convolution Theorem

CS 4495 Computer Vision – A. Bobick Frequency and Fourier Transform

An Example of Convolution

Mirroring h
about the origin

Translating the
mirrored
function by x

Computing the
sum for each x
CS 4495 Computer Vision – A. Bobick Frequency and Fourier Transform

An Example of Convolution

It causes the
wraparound
error

It can be
solved by
appending
zeros
CS 4495 Computer Vision – A. Bobick Frequency and Fourier Transform

Zero Padding

► Consider two functions f(x) and h(x) composed of A and B

samples, respectively

► Append zeros to both functions so that they have the same

length, denoted by P, then wraparound is avoided by choosing

P ≥A+B-1
CS 4495 Computer Vision – A. Bobick Frequency and Fourier Transform

Zero Padding

► Let f(x,y) and h(x,y) be two image arrays of sizes A×B and C×D
pixels, respectively. Wraparound error in their convolution can
be avoided by padding these functions with zeros
CS 4495 Computer Vision – A. Bobick Frequency and Fourier Transform

Summary
CS 4495 Computer Vision – A. Bobick Frequency and Fourier Transform

Summary
CS 4495 Computer Vision – A. Bobick Frequency and Fourier Transform

Summary
CS 4495 Computer Vision – A. Bobick Frequency and Fourier Transform

Summary