Motion estimation plays an important role in the video encoding process. Computational complexity and motion estimation time are the main factors which affect the performance of the video encoding process. The block-based motion estimation algorithms are the widely used methodology and its search patterns have more impact on searching speed and distortion performance in the video encoding process. In this paper, the adaptive hexagonal search (AHS) based motion estimation algorithm is proposed for efficient video encoding systems with the help of two hexagon shaped search patterns such as Large Hexagon Search Pattern (LHSP) and Small Hexagon Search Pattern (SHSP). The search process of AHS initiates from the center of the search window using LHSP which is continuously utilized until the center position of LHSP satisfies the minimum bounding distortion (MBD). The search pattern is changed from LHSP to SHSP when the position of MBD within the boundary of LHSP. During this process, the obtained MBD point is considered an optimal solution for the motion vector. Finally, the motion vector is obtained based on the MBD from the SHSP. The adaptive hexagonal search based motion estimation is implemented in the MATLAB version 2018 software tool. The proposed AHS methodology is compared with the FSA, DS, ABC-BM algorithms in terms of PSNR and the experimental results showed that the proposed AHS based motion estimation improved the performance at averagely 24.27% in terms of PSNR.
Volume 11 | 02-Special Issue
Pages: 105-114