Single-Dimension Perturbation Glowworm Swarm Optimization Algorithm for Block Motion Estimation

Algorithm 2

Single-dimension perturbation glowworm swarm optimization algorithm (SDGSO).

Step  1.   Read the data.

Step  2.   Set parameters: 𝜌, 𝛾, 𝛽, 𝑛_𝑡, 𝑠_min, 𝑠_max, 𝑎.

Step  3.    Set the size of block 𝑚𝑏𝑆𝑖𝑧𝑒 and divide the current frame image into 𝑧 rules of

block.

Step  4.    for 𝑘=1 to 𝑧 do

     (4.1)  Predict the MV of the current macro block and the center of search window.

     (4.2)  Generate initial population of glowworms 𝑥_𝑖(𝑖=1,2,…,𝑛)

      (4.3)  for each 𝑖=1 to 𝑛 initializing the luciferin value l(i);

     (4.4)   Set 𝑡=1;

     (4.5) for each (𝑡≤max𝑡) doing

    (4.5.1) for each glowworm i doing

     (4.5.1.1) Form the neighborhood 𝑁_𝑖(𝑡);

         (4.5.1.2) for each glowworm 𝑗∈𝑁_𝑖(𝑡), computing probability 𝑃_𝑖𝑗(𝑡)

according

          to the formula (3);

     (4.5.1.3) Select glowworm 𝑗 using 𝑃_𝑖𝑗(𝑡);

     (4.5.1.4) Update glowworm step s with the formula (8);

     (4.5.1.5) Update glowworm position 𝑥_𝑖 with the formula (2);

     (4.5.1.6) Search the new location by using SDPS;

     (4.5.1.7) Update the luciferin value 𝑙(𝑖) according to the formula (8);

     (4.5.1.8) Update local-decision domain 𝑟^𝑖_𝑑(𝑡+1) according to the formula (4);

     (4.5.1.9) if 𝑥_𝑖 keep unchanged for limt generation then terminating the

iteration;

Step  5. Output MV.