Performance Analysis of Fast Block Matching Algorithms in Motion Estimation

Redzuan, Abdul Manap (2009) Performance Analysis of Fast Block Matching Algorithms in Motion Estimation. Project Report. UTeM, Melaka, Malaysia. (Submitted)

PDF (Full Text) Performance_Analysis_of_Fast_Block_Matching_Algorithms_in_Motion_Estimation_-_Redzuan_bin_Abdul_Manap.pdf - Submitted Version Restricted to Registered users only Download (3MB)

Abstract

Block matching algorithm (BMA) techniques have been widely used as a method to achieve high compression ratio in video applications. In recent years, many of these BMAs have been developed with similar intention of reducing the computational costs and requirements while at the same time maintaining the video signal quality. The main objective of this project is to determine the best BMA to be implemented based on the video sequence category. This is done by first reviewing the video compression techniques and the BMA in general, the available BMAs, their advantages and disadvantages before implementing them onto various type of standard test video sequence using MATLAB. For this project, several well-known fast BMAs, namely Three Step Search (TSS), New Three Step Search (NTSS), Four Step Search (4SS) and Diamond Search (DS) are selected for implementation. The performance of these BMAs are then analyzed and compared in terms of peak signal-to-noise ratio (PSNR), number of search points needed as well as their computational complexity. Simulation results demonstrate that OS algorithm has reduced other algorithm's computational work up to 48% on average while NTSS algorithm performs the best in terms of PSNR analysis with up to 1.884 dB improvement on video quality.

Actions (login required)

View Item