Compression of synthetic images

Although the "digital convergence" (of computers and video) is well underway, the convergence of computer graphics and video technology has only begun. A key problem in this area is developing efficient ways to compress the data used in graphics algorithms: geometric models, synthetic images, textures, and volumes. Among the questions we are addressing in this research are:

• Can a rendering engine provide hints to a compression engine? Examples might include motion hints, quantization hints, importance hints, or real-time constraints.
• Can a rendering engine perform compression or decompression? More specifically, can we render directly from compressed textures or volumes, or can we render directly to compressed image formats?
• What compression algorithms can we invent that are specialized for compressing geometry and synthetic images? Examples might include specialized basis functions for compressing smoothly shaded images and hybrids of geometry-based and image-based compression.

This project is currently inactive.

People:

• Maneesh Agrawala <maneesh@pepper.stanford.edu >
• Andrew Beers < beers@cs.stanford.edu >
• Navin Chaddha
• Marc Levoy

Recent papers in this area:

Polygon-Assisted JPEG and MPEG Compression of Synthetic Images

Marc Levoy

Proc. Siggraph '95

Model-Based Motion Estimation for Synthetic Animations

Maneesh Agrawala, Andrew Beers and Navin Chaddha

Proc. ACM Multimedia '95

Rendering from Compressed Textures

Andrew Beers, Maneesh Agrawala and Navin Chaddha

Proc. SIGGRAPH '96

A list of technical papers, with abstracts and pointers to additional information, is also available. Or you can return to the research projects page or our home page.