Performance Analysis of Motion Estimation Algorithms on Digital Video Images

Abstract

An improvement on redundancy to achieve high compression ratio in video coding is developed. Block Matching Motion Estimation (BMME) techniques have been particularly used in various coding standards. In the BMME, search patterns with different shapes or sizes and the center-biased characteristics of motion vector (MV) have large impact on the search speed (search points) and peak signal-to-noise ratio (PSNR) as the quality of video images. These basic algorithms are Full Search and other two fast search methods. The Cross Diamond Search (CDS) algorithm was designed to fit the cross-center-biased (CCB) MV distribution characteristics of the real-world video sequences. CDS compares favorably with the other algorithms for low motion sequences in terms of speed, quality and computational complexity.

Adaptive Prediction based Diamond Search Algorithm (APDS) for Video Encoder of Solar Powered UAV

“…SizeOfPSAP Max MV |X |,| Y |   (1) where X and Y are the coordinates of MV of previous MB.…”

“…Where accurate match is found, difference in the positions of current and reference frame is calculated which is called Sum of Absolute Difference (SAD). The displacement vector for this SAD is defined as the Motion Vector (MV) [1][2]. The current frame is divided into nonoverlapping blocks like 16×16, 16×8, 8×16, 8×8 etc.…”

Performance Analysis of Fast Block Matching Motion Estimation Algorithms

“…Where accurate match is found, difference in the positions of current and reference frame is calculated which is called Sum of Absolute Difference (SAD). The displacement vector for this SAD is defined as the Motion Vector (MV) [1][2].…”