Are there any "special" image compression algorithms for face cases?

Are there any "special开发者_StackOverflow中文版" image compression algorithms for face cases? So i'm creating a conference programm I want to transfer images (or videos) of talking heads through the internet. Are there any special algorithms to compress images/videos of talking heads so to make them smaller (like special voice compression algorithms)?

Really not my field, but I know that a couple of researchers from my university work on a compression algorithm that shows impressive results on limited domains (such as faces). They have published a few articles explaining the algorithms, as well as released a Matlab plugin for it.

The algorithm is called K-SVD, and you can read more about it in the basic article. There are many follow-up articles published as well.

For the implementation, check out:

• Prof. Elad's Software page
• Prof. Rubinstein Software Page

Jörgen Ahlberg at Linköping University, Sweden, wrote a PhD thesis on the subject in 2002:

J. Ahlberg, Model-based Coding - Extraction, Coding, and Evaluation of Face Model Parameters , PhD Thesis No. 761, Dept. of Electrical Engineering, Linköping University, Sweden, 2002.

It appears the technology has been moved to http://www.visagetechnologies.com/

Edit
The algorithm Jören developed is called Candide-3, more info on the algorithm is available at http://www.icg.isy.liu.se/candide/