ECE160 - CMPSC182
Assignment 6, Spring 2008
Due: In class, 5p.m. May 19


1.  Image compression algorithms use a combination of:
  Encoding to separate Luminance and Chrominance
  Discrete Cosine Transformation
  Quantization
  Zigzag Ordering
  Run length coding
  Entropy coding.

Explain how these techniques work together and how each 
of the techniques is necessary for the others to be effective.

2.  The logarithmic motion vector search method depends on a form
of linearity in the image to be effective.  Explain why the algorithm
requires that linearity and justify the assumption for typical images.
Suggest a kind of image for which the assumption is not justified and
for which the algorithm might fail.

Does the hierarchical motion vector search algorithm depend on a similar
assumption?  Can you suggest a kind of image for which the hierarchical
algorithm might fail?