Evaluation and comparison of different fast block matching algorithms using motion estimation

This paper presents a different types of block matching algorithms using motion estimation are evaluated where the effects of the macro block size will be reviewed that range from the very basic Full Search Algorithm to the recent Adaptive Rood Pattern Search Algorithm. Seven different types of block matching algorithms are implemented. Each algorithm is evaluated using different video sequences and comparisons are made through computation time and peak signal noise ratio (PSNR). The simulation results showed that the ARPS algorithm is very effective in reducing the computations time and achieves very good peak signal to noise ratio values. Thus ARPS algorithm is more efficient than the conventional searching algorithms for video compression standard. The algorithms that are evaluated in this paper are widely accepted by the video compressing community and have been used in implementing various standards, ranging from MPEG1 / H.261 to MPEG4 / H.263.