A novel Cooperative Motion Estimation Algorithm based on Particle Swarm Optimization and its multicore implementation

作者：

Highlights：

• Collaboration between neighboring MBs is allowed during motion estimation.

• Cooperation between neighboring MBs allows refining motion vectors found so far.

• Temporal correlation used in PSO initialization removes dependency between MBs.

• The maximum allowed velocity of the PSO particles is adaptively varied.

• An efficient multi-core implementation using Matlab is presented.

摘要

•Collaboration between neighboring MBs is allowed during motion estimation.•Cooperation between neighboring MBs allows refining motion vectors found so far.•Temporal correlation used in PSO initialization removes dependency between MBs.•The maximum allowed velocity of the PSO particles is adaptively varied.•An efficient multi-core implementation using Matlab is presented.

论文关键词：Motion estimation,Cooperative multi-swarm,PSO,Parallel processing

论文评审过程：Received 3 November 2014, Revised 11 September 2015, Accepted 11 September 2015, Available online 25 September 2015, Version of Record 3 October 2015.

论文官网地址：https://doi.org/10.1016/j.image.2015.09.010