Title :
Foveated online 3D visualization
Author :
Basu, Anup ; Cheng, Irene ; Pan, Yixin
Author_Institution :
Dept. of Comput. Sci., Alberta Univ., Edmonton, Alta., Canada
Abstract :
Spatially varying sensing (foveation) was first used as a means for image compression in our past research (1993). In this report we extend previous work by us to address the advantages of foveation in improving the performance of interactive 3D visualization over bandwidth limited channels, such as the Internet. We develop an algorithm that is implemented in JAVA3D. The method combines foveating JPEG texture files with level-of-detail representation in JAVA3D. Applications of the work include model based coding following the Synthetic-Natural Hybrid (SNHC) paradigm in MPEG-4 to Electronic Commerce with 3D visualization of merchandise. Experimental results are presented to validate our approach.
Keywords :
Internet; data compression; image coding; multimedia communication; Electronic Commerce; Internet; JAVA3D; JPEG texture files; MPEG-4; Synthetic-Natural Hybrid paradigm; bandwidth limited channels; foveation; image compression; interactive 3D visualization; level-of-detail representation; merchandise; model based coding; spatially varying sensing; Bandwidth; Image coding; Image resolution; Image storage; Internet; Java; MPEG 4 Standard; Ray tracing; Transform coding; Visualization;
Conference_Titel :
Pattern Recognition, 2002. Proceedings. 16th International Conference on
Print_ISBN :
0-7695-1695-X
DOI :
10.1109/ICPR.2002.1048193