EECE 545: Signal Compression

Administrative Information - Fall 1999

Course Outline
  1. Introduction (1 week)
    • Why compression?
    • Compression techniques: lossless and lossy
    • Data and signal types: computer files, text, speech, image, video
    • Applications: telephony, facsimile, CD-ROM, digital TV, WWW
       
  2. Lossless Compression (3 weeks)
    • Information theory: Basic concepts
    • Minimum redundancy coding: Huffman coding
    • Arithmetic coding
    • Modeling: pdf, prediction
    • Dictionary techniques
       
  3. Lossy Compression (5 weeks)
    • Rate-distortion theory
    • Distortion: Auditory perception, HVS, mathematical measures
    • Scalar quantization
    • Vector quantization, structures
    • Predictive coding
    • Transform coding
    • Subband/wavelet coding
    • Analysis/synthesis and modeling techniques
    • Motion estimation/compensation
       
  4. Applications (3 weeks)
    • Requirements: compression, quality, complexity, delay, sensitivity
    • Standards: LZW, G.722, G3/G4, JBIG, JPEG, H.261/H.263, MPEG
       
Number of Lecture Hours:  3-0-0
 

Text: Introduction to Data Compression, by K. Sayood.
 

Grading

  • Homework: 10%
  • Midterm exam 1: 30%
  • Midterm exam 2: 30%
  • Project: 30% (10% presentation, 10% demo, 10% written report).
Course schedule:
 
  Exams Homework Project
16 Sep Lecture 1      
23 Sep Lecture 2      
30 Sep  Lecture 3      
7 Oct Lecture 4      
14 Oct Lecture 5   Hw #1 due 18 Oct in my office. Solutions available Oct 19.  Written project proposal due with Hw #1.
21 Oct Lecture 6  Quiz #1     
28 Oct Lecture 7       
4 Nov Lecture 8      
11 Nov Remembrance Day      
18 Nov Lecture 9       
25 Nov Lecture 10   Hw #2 due.  
2 Nov Lecture 11 Quiz #2    
6 Dec     Project presentations. Written reports due.

Miscellaneous

A Mathematical Theory of Communication by Claude E. Shannon.
 
 

Fall 1999 Project Presentations