Nature inspired algorithms (NIA) for efficient video compression – A brief study

Abstract Motion Estimation in a video is a challenging optimization problem where the objective is to minimize the error or to maximize the correlation between a Macro Block (MB) in current frame and MB in the reference frame and to find the best matching MB. As the pace of development of conventional Pattern-Based block matching algorithms was very fast and also so many in numbers that it almost got saturated. But at the same time rapid development of new algorithms based on natural genetics to exploit intelligence such as genetic algorithms, evolutionary algorithm, particle swarm optimization and differential evolution show the potential of these algorithms in optimization and gave researchers a broad dimension to apply such nature inspired algorithms in the field of motion estimation. Subsequently, researchers started implementing one by one many nature-inspired algorithms, sometimes referred to as soft computing techniques, for solving the optimization problem of motion estimation and proved that soft computing techniques have huge advantages over the conventional pattern-based or prediction based block matching algorithms. In this paper, several nature-inspired algorithms that are implemented for video motion estimation are reviewed and the performance is compared to highlight the competitive advantages of soft computing techniques over existing fixed pattern search algorithms.