Tutorials

Three Tutorials
Click on the titles for more details

[ Time Table Main Page ]


 

Multimedia Streaming: MPEG-4 approach

Wael Badawy
University of Calgary
Canada

Overview and Objectives:
MPEG-4 is a new ISO/IEC standard that targets streaming multimedia. The MPEG (Moving Picture Experts Group) developed MPEG-1 and MPEG-2 that target interactive video on CD-ROM and Digital Television, respectively. The MPEG-4 provides tools to deliver multimedia content over different communication channels and targets a wide range of interactive applications. MPEG-4 provides new features such as scalability which enable the providers to transmit the same content using different channels, error-resilient techniques that allow robust transmission over wireless transmission, content-based approach that allows higher levels of interaction with scene contents. The MPEG-4 scene is presented as aural, visual or audiovisual objects or AVOs. The MPEG-4 became an International Standard for multimedia streaming in January 1999 and its second version, as an amendment to Version 1, achieved the status of 'Final Draft International Standard' in December 1999.

This tutorial will introduce the MPEG-4 Version 1 and Version 2 visual coding tools and their functionalities. It will focus on the applications of MPEG-4 for multimedia streaming and how wired and wireless multimedia streaming can benefit from MPEG-4 visual coding tools.

Topics:
The tutorial topics:

  • Brief introduction to Multimedia Structure
  • MPEG-4 scope and structure
  • MPEG-4 Version 1 and 2 tools
  • Visual
  • Audio
  • Systems
  • MPEG-4 Profiles
  • Streaming Multimedia using MPEG-4
  • M4IF
  • Wireless Multimedia Forum
  • Demo and Summary

The tutorial will target audience from several backgrounds.

Instructor Biography:
Wael Badawy received his Ph.D. from the department of Computer Engineering, University of Louisiana at Lafayette, LA and his M. Sc. in Computer Engineering from the same university. He also had an M. Sc. in Computer Science from Alexandria University, 1997 and B. Sc. in Computer Science from Alexandria University in 1994. Dr. Badawy's research interest include video coding for low-bit rate applications, digital video processing, video library, watermarking, spatial database, low power, design Methodologies, microelectronics, VLSI and CAD tools. His research involves designing new models, techniques, algorithms, architectures and low power prototype for MPEG-4 consumer products. Dr. Badawy was honored with the "1998 Upsilon Pi Epsilon Honor Society and IEEE Computer Society Award for Academic Excellence in Computer Disciplines"

Top

It is Time to Fuzzify Artificial Neural Networks

Ajith Abraham
Gippsland School of Computing & Information Technology, Monash University, Australia

Neural Networks and fuzzy inference systems have been widely used in several intelligent multimedia applications. Artificial Neural Network learns from scratch by adjusting the interconnections between layers. A valuable property of neural network is that of generalization, whereby a trained network is able to provide a correct matching in the form of output data for a set of previously unseen input data. Fuzzy Inference System is a popular computing framework based on the concept of fuzzy set theory, fuzzy if-then rules, and fuzzy reasoning. With crisp inputs and outputs, fuzzy inference system implements a nonlinear mapping from its input space to output space by a number of if-then rules.

Integrating Neural Networks and Fuzzy Inference Systems have attracted the growing interest of researchers in various multimedia applications due to the growing need of adaptive intelligent systems to meet the real world requirements. There are several approaches to integrate neural networks and fuzzy inference systems and very often it depends on the application. We broadly classify the integration of neuro fuzzy systems into three categories namely concurrent model, cooperative model and fully integrated model. We briefly discuss the features of each model and generalize the advantages and deficiencies of each model. We will also attempt to give some insights when to use which model.

This tutorial starts with some basic theoretical aspects of neural networks and fuzzy inference systems and their application areas stressing the advantages of each technique. We further discuss the step-by-step modeling of different neuro-fuzzy architectures. We also demonstrate how neuro-fuzzy techniques could be used for many practical applications involving image classification and data mining tasks etc.

Topics:

  • What is soft computing?
  • From mammalian neuron to artificial neural networks
  • Different neural network learning algorithms
  • Feedforward neural networks
  • Kohenen's self organizing maps
  • Advantages of neural networks
  • Fuzzy inference system
  • Mamdani fuzzy inference system
  • Takagi Sugeno fuzzy inference system
  • Advantages of fuzzy inference systems
  • Why we need neuro-fuzzy systems?
  • Types of neuro-fuzzy systems
  • Some popular neuro-fuzzy architectures: ANFIS, EFuNN, NEFCLASS others
  • Advantages of different neuro-fuzzy models
  • Application areas with simple examples
  • Beyond neuro-fuzzy: perspectives and directions

Instructor Biography:
Ajith Abraham is a research scholar in GSCIT, Australia. He received MS degree in computer control and automation from Nanyang Technological University, Singapore and Bachelors degree in EEE with first class (Hons) from the Calicut University, India. He has 8 years of Industrial experience. and also a member of IEEE & IEEE computer society, ACM, Institution of Electrical Engineers (UK) and Institution of Engineers (Australia).

His current research interests are in development hybridization of intelligent systems, neural networks, fuzzy inference systems, global optimization algorithms, scheduling techniques and parallel computing. He is interested both in system level development as well as industrial applications.

Top

Image and Video Compression Techniques and Standards

S.R.Subramanya
University of Missouri, Rolla
USA

Overview and Objectives:

The phenomenal increases in the generation, transmission, and use of digital images and video in many applications is placing enormous demands on the storage space and communication bandwidth. Data compression is a viable approach to alleviate the above storage and bandwidth demands. This tutorial is intended to give an insight into a few major image and video compression techniques and a brief look at the popular image and video coding standards.

The tutorial starts with an overview of the basics of lossy compression techniques in general. The concepts and issues in quantization are introduced. Predictive image coding techniques are then covered, which primarily includes DPCM and ADPCM. Transform coding techniques using DFT and DCT are then addressed. A brief overview of Subband coding is then given, followed by Vector Quantization. The bi-tonal image compression standard JBIG, and continuous-tone image standard JPEG are then covered. Video coding techniques is the next topic which covers elements of motion estimation and motion compensation. This is followed by the motion video standard MPEG, and videoconferencing/videotelephony standards H.261/263.

This tutorial enables the participant to:

* understand the principles of lossy data compression.
* get an insight into a few of the major image and video coding techniques.
* identify the issues involved and the evaluation metrics.
* get an overview of the common image and video compression standards.

Topics Covered:

Lossy Image Compression Basics, Scalar Quantization
Predictive coding (DPCM, ADPCM)
Transform Coding (FFT, DCT)
Subband Coding
Vector Quantization
JBIG, JPEG Standards
Video Coding Techniques
Video Coding Standards (MPEG, H.261/263)

Biography:
S.R.Subramanya received his doctoral degree in computer science from George Washington University, Washington, D.C. and masters degree in computer science from Indiana University, Bloomington. He was the recipient of Richard Merwin memorial award at George Washington University in 1996, and Grant-In-Aid of Research award from Sigma-Xi in 1997. He is currently an Assistant Professor at the University of Missouri, Rolla. He has been teaching courses in Data Compression and Multimedia Information Systems. His research involves Multimedia data classification, indexing, and retrieval, Data Compression, and Parallel algorithms and architectures. He has served as a reviewer, program committee member, and session chair of several International Conferences. He has also been a reviewer for several journals.

Top

Return to Home Page