Improved Block Matching Motion Estimation Technique using Modified Particle Swarm Optimization in Video Coding

Abstract

In video coding, the block matching based motion estimation (BMME) schemes by exploiting the uni-modal error surface is quite popular for easy implementation. Though, many real-world video sequences exhibit multiple local minima within the block search window. This research article proposes a pattern based modified particle swarm optimization approach for motion estimation (PMPSO-ME). The population-based evolutionary method; PSO ensures the global optimum solution and avoids trapping into local minima.The proposed technique is analyzed on JM 18.6 reference software of H.264/AVC platform. The proposed technique is evaluated for compared with the state-of-the-art approaches in terms of number of search points, encoding time complexity, Bjontegaard metrics etc.