Figure 1:   Effect of error propagation
\begin{figure}
\vspace{0.3in}
\centerline{
\psfig {figure=errorprop2.eps,height=3in,width=5in}
}
\vspace{0.06in}
{\bf }
\vspace{0.1in}\end{figure}

It has been projected that digital multimedia products will have a $10 billion market in the next five years. Especially, state-of-the-art image/video compression for video-conferencing, HDTV, video telephony etc. based on international standards like JPEG, MPEG, and H.263 will have a major share in this market. Companies and universities all over the world are gearing themselves to design and develop new algorithms for low bit-rate, error-resilient transmission of video. Recently, companies like Lucent Technologies, Panasonic, IBM, SUN, Motorola, Hewlett Packard and many others have started a digital video department to design various digital video algorithms and products.

The objective of this course is to introduce students (both graduate and undergraduate) to the current techniques in image/video compression. Basic theory behind these methods such as information theory, time domain and transform domain (DCT, wavelets) based compression, joint source-channel coding for video transmission over noisy channels, and digital video security will be introduced.

JPEG, MPEG, and H.263 standards will be studied. Applications to HDTV, internet movies, and video-conferencing will be discussed. Recent trends in digital image and video watermarking, data hiding and watermark attack will also be introduced. Students will form groups and implement image/video compression algorithms and perform various experiments.

Introductory knowledge of probability will be assumed. Computer programming knowledge using C and MATLAB is required. Both graduate and senior undergraduate students who satisfy the above requirements can register for this course. Others should talk to the instructor before registering.

If you have a problem with the schedule, let me know on the first day of class or send me an email.



  Main Page
ECE