CS 7321: Computer Vision 1

Problem Set 2: Images and Spatial Properties

By Jiqun Wang

Goal: This PS is designed to help you use matlab and also learn about Spatial Properties of Images.


 

Problem:

All of this relates to the HANDOUT from Computer Vision by Ballard and Brown, Prentice-Hall, 1982, Section 2.2.4. Handed out in Class Tue Feb 10, 1998.
  1. Original image:stripes.jpg
  2. Generate images as shown by Figure 2.5 (b), (c), (d), (e) and (f).

    3. ( Scanned Image of that page )
    Caption: Fig 2.5
    (a) An image, f(x,y),
    (b) A rotated version of (a),filtered to enhance the high spatial frequencies.
    (c) Similar to (b), but but filtered to enhance low spatial frequencies.
    (d), (e) and (f) show the logarithim of the power spectrum of (a), (b) and (c). The power spectrum is the log square modulus of the Fourier Transform F(u,v). Considered in polar coordinates (rho,theta), points of small rho correspond to low spatial frequencies ("slow varying" intensities), large rho to high spatial frequencies contributed by "fast" variations such as step edges. The power at (rho, theta) is determined by the amount of intensity variation at the frequency of rho occuring at angle theta.
  3. Experiment with a few other synthetic and real images.
    4. Original Image Operations Output Image
     operation #1:
    A rotated version of (a),filtered to enhance
    the high spatial frequencies.
      operation #2:
    (c) is Similar to (b), but but filtered to
    enhance low spatial frequencies.
      operation #3:
    (d) show the logarithim of the power spectrum
    of (a)
      operation #4:
    (e) show the logarithim of the power spectrum
    of (b)
      operation #5:
    (f) show the logarithim of the power spectrum
    of (c)
    a new image for practice high pass filtered low pass filtered
    fft2 of above fft2 of above fft2 of above
Source Code