A “Tuned” Mask Learnt Approach Based on Gravitational Search Algorithm

Pseudocode 2

Pseudocode of learning the “Tuned” mask based on GSA.

Begin

Input training sample texture images

Set the parameters of GSA and generate initial populations

For each agent (object), generate a “Tuned” mask by using (15) (the position of the agent could be directly used as the

element value of the mask), make convolution with training images and “Tuned” mask, and output the eigenvalues

While (The current iteration t < The maximum iteration T)

Compute the fitness value of each object by using (16)

Update the gravitational variable , and and of the population

Calculate the active gravitational mass , the passive gravitational mass , the inertial mass and the

acceleration for each object

Update velocity and position of each object by using (6)

If (The fitness value of current position is better)

Replace the object by the new position

End if

End while

Output the optimal “Tuned” mask according to (15)

End