Share Email Print

Proceedings Paper

Proxy-based approaches for IDCT acceleration
Author(s): Wendi Pan; Antonio Ortega; Ibrahim N. Hajj-Ahmad; Roberto Sannino
Format Member Price Non-Member Price
PDF $17.00 $21.00

Paper Abstract

In the boundary between wired and wireless worlds, proxy-based processing has been recognized as an efficient approach to provide client-specific adaptation to improve the overall performance. For the portable video decoding devices with wireless access ability, fast, low-power decoding is desirable. In this paper, we propose a novel proxy-based framework that is capable of accelerating the IDCT operations performed at the client. We introduce a proxy-specific IDCT algorithm that is not only suitable for the proxy framework, but has fine-grained complexity scalability as well. We then demonstrate the effectiveness of the proxy by two examples. The first example is a simple proxy that is assigned the task of IDCT block classification, which is performed at the client along with the IDCT operation conventionally. Experimental results clearly show the complexity advantage of this method. The second example is a more active proxy. Adaptation to the image statistics is introduced. Simulations of the optimization process based on the average complexity criteria show that the client complexity can be further reduced, while maintaining a low side information overhead. Both examples provide the trade-off between the client-proxy bandwidth increment and the IDCT complexity reduction at the client.

Paper Details

Date Published: 29 December 2000
PDF: 12 pages
Proc. SPIE 4310, Visual Communications and Image Processing 2001, (29 December 2000); doi: 10.1117/12.411841
Show Author Affiliations
Wendi Pan, Univ. of Southern California and ST Microelectronics (United States)
Antonio Ortega, Univ. of Southern California (United States)
Ibrahim N. Hajj-Ahmad, ST Microelectronics (United States)
Roberto Sannino, ST Microelectronics (United States)

Published in SPIE Proceedings Vol. 4310:
Visual Communications and Image Processing 2001
Bernd Girod; Charles A. Bouman; Eckehard G. Steinbach, Editor(s)

© SPIE. Terms of Use
Back to Top