Share Email Print
cover

Proceedings Paper

Scalable complexity-distortion model for fast motion estimation
Author(s): Xiaoquan Yi; Nam Ling
Format Member Price Non-Member Price
PDF $14.40 $18.00
cover GOOD NEWS! Your organization subscribes to the SPIE Digital Library. You may be able to download this paper for free. Check Access

Paper Abstract

Recently established international video coding standard H.264/AVC and the upcoming standard on scalable video coding (SVC) bring part of the solution to high compression ratio requirement and heterogeneity requirement. However, these algorithms have unbearable complexities for real-time encoding. Therefore, there is an important challenge to reduce encoding complexity, preferably in a scalable manner. Motion estimation and motion compensation techniques provide significant coding gain but are the most time-intensive parts in an encoder system. They present tremendous research challenges to design a flexible, rate-distortion optimized, yet computationally efficient encoder, especially for various applications. In this paper, we present a scalable motion estimation framework for complexitydistortion consideration. We propose a new progressive initial search (PIS) method to generate an accurate initial search point, followed by a fast search method, which can greatly benefit from the tighter bounds of the PIS. Such approach offers not only significant speedup but also an optimal distortion performance for a given complexity constrain. We analyze the relationship between computational complexity and distortion (C-D) through probabilistic distance measure extending from the complexity and distortion theory. A configurable complexity quantization parameter (Q) is introduced. Simulation results demonstrate that the proposed scalable complexity-distortion framework enables video encoder to conveniently adjust its complexity while providing best possible services.

Paper Details

Date Published: 24 June 2005
PDF: 11 pages
Proc. SPIE 5960, Visual Communications and Image Processing 2005, 59603Y (24 June 2005); doi: 10.1117/12.632687
Show Author Affiliations
Xiaoquan Yi, Santa Clara Univ. (United States)
Nam Ling, Santa Clara Univ. (United States)


Published in SPIE Proceedings Vol. 5960:
Visual Communications and Image Processing 2005
Shipeng Li; Fernando Pereira; Heung-Yeung Shum; Andrew G. Tescher, Editor(s)

© SPIE. Terms of Use
Back to Top