Abstract of the research project 1. Aim Today's modern multimedia applications have risen to a level of quality (see the graphical detail of virtual threedimensional scenes, or the HDTV or even cinema resolution of video content) that would have been considered unattainable until recently. Communications have also seen an explosion in terms of network and user heterogeneity, ranging from ultra-fast optical links to wireless error-prone channel mediums, and from low-power mobile devices to highend desktop computers. Delivering multimedia content in such rapidly-changing heterogeneous environments requires new types of compression systems that simultaneously provide scalability and resilience against transmission errors, and enable an optimal, on-the-fly content adaptation for the given network conditions and users characteristics. 2. Objectives Basically, the project continues the major objectives and research activities set in the first term of our FWO mandate. In this context, the current results provided by geometric transforms are very promising, as they significantly outperform the classical wavelet bases in the context of still image coding. These findings motivate the quest for finding new geometric bases that further improve the performance against that of classical wavelet bases, or that are applicable in other domains such as video coding, or coding of 3D graphics. Designing new geometric representations for images, video and 3D graphics, and developing new scalable error-resilient compression systems that are based on such representations are thus the two major objectives in our project.

Runtime: 2005 - 2007