DocumentCode :
571825
Title :
UMHexagonS based motion estimation architecture comparison
Author :
Juri, Arief Affendi ; Jambek, Asral Bahari
Author_Institution :
Sch. of Microelectron. Eng., UniMAP, Kangar, Malaysia
Volume :
1
fYear :
2012
fDate :
12-14 June 2012
Firstpage :
374
Lastpage :
377
Abstract :
This paper present the existing design and review of hardware for UMHexagonS based motion estimation for H.264/AVC video compression. Four existing motion estimation architecture that implement UMHexagonS are presented, analysed and compared to show the area to tackle for future improvements. The presented architectures are also compared against our proposed architecture. The results are compared in term of gate count and throughput with emphasize on throughput as it shows the speed of the architecture. The proposed architecture gives a promising result compared with some of the reviewed architecture. It shows an improvement of 91% compared to the full search architecture and 20% compared to a basic UMHexagonS architecture.
Keywords :
data compression; motion estimation; UMHexagonS; full search architecture; future improvements; motion estimation architecture; reviewed architecture; video compression; Adders; Computer architecture; Motion estimation; Random access memory; Standards; Throughput; Video coding; H.246/AVC; Motion Estimation; UMHexagonS; architecture;
fLanguage :
English
Publisher :
ieee
Conference_Titel :
Intelligent and Advanced Systems (ICIAS), 2012 4th International Conference on
Conference_Location :
Kuala Lumpur
Print_ISBN :
978-1-4577-1968-4
Type :
conf
DOI :
10.1109/ICIAS.2012.6306221
Filename :
6306221
Link To Document :
بازگشت