Elec 695: Image and Video Communications
Spring 1998 - Rice University
|
Instructor
|
Aria Nosratinia, 2019
Duncan Hall, Ext. 5056, aria@ece.rice.edu
|
|
Time
|
Tue-Thu, 2:30-3:50pm
|
|
Place
|
Room 117 Physics Lab
|
|
Text
|
Course notes, journal articles, and other material that will be
distributed as necessary.
|
|
References
|
Cover & Thomas, Elements of Information Theory, John Wiley
Tekalp, Digital Video Processing, Prentice-Hall
Gersho & Gray, Vector Quantization and Signal Compression, Kluwer
|
|
Grading
|
Midterm exam (%20), Project report (%60), Homeworks (%10), Class participation
(%10)
|
|
Prerequisite
|
One semester of advanced random processes, one-semester background in
Information theory. Elec 540 (Source Coding and Compression) is helpful
but is not required.
|
|
Office Hours
|
TBA
|
The booming growth of communication technologies, especially in
networking and wireless, has given rise to the promise of many new and
exciting applications. Image and video communications are among the most
rapidly expanding sections of this boom. Applications of image and video
communication range from computers and multi-media to videophones and
wireless personal communication systems (PCS), from High-Definition
Television (HDTV) to entertainment applications such as the Digital
Video Disc (DVD). In terms of both technical and theoretical challenges
as well as economic promise, this is a most unique and exciting area.
This course explores the theory and practice of image and video
communications. A course in probability and random processes and a
one-semester background in information theory is prerequisite. Elec540,
``Source Coding and Compression,'' is a helpful background but is not
required.
Elec695 consists of lectures, as well as a project. The project will be
chosen and carried out in consultation with the instructor. For the
subject of the project, the students have a choice to either survey an
existing body of work, or explore new territory. In either case, the
students will analyze methods, write computer programs to perform
experiments, and report their results.
This course is designed to develop an in-depth understanding of the
subject, based on first principles. Thus it can provide an entry point
for those who may want to pursue research in this area.
Contents:
The following is a tentative syllabus of Elec695.
- Overview
- Fundamentals of Information Theory
- Entropy Coding
- Scalar Quantization
- Vector Quantization
- Images and video
- Sampling
- Digital representations
- Overview of video signals
- Image perception
- Image Compression
- Transform Coding
- Bit allocation
- The Karhunen-Loeve Transform
- Discrete Cosine Transform
- The JPEG standard
- Wavelet Image Compression
- Subband and Wavelet transforms
- Energy compaction of wavelets
- Zerotree wavelet coding
- Prediction and vector quantization in subbands
- Trellis Coded Quantization (TCQ)
- Lagrangian methods and R-D optimization
- Space-frequency and wavelet packets
- Motion Estimation
- Role of motion in video coding
- Block motion estimation - block matching
- Overlapped-block motion estimation
- Warping motion estimation
- Pixel-based methods
- Video Coding
- Quantization/motion tradeoff
- The rate control problem
- Videophone - H.261 and H.263 standards
- The MPEG-1 Standard
- The MPEG-2 Standard
- Model-based video coding
- Advanced Topics in Image and Video Coding
- Joint warping overlapped-block motion estimation
- Rate-distortion optimized video coding
- Progressive (embedded) coding of video
- Joint source-channel coding
- Post processing
- Projection on convex sets
- Wavelet filtering and interpolation
- Error concealment in image transmission
- Error concealment in video transmission
Aria Nosratinia
Last modified: January 12, 1998