Painting Using Genetic Algorithm with Aesthetic Evaluation of Visual Quality - Technologies and Applications of Artificial Intelligence - page 131

(a) The better image that has a smaller bounding box

(b) The worse image that has a larger bounding box

Fig. 5. Laplacian filter example

this ratio as the first part of fitness evaluation and compute it by

W b H b
WH ,

f edge =1 −

where W b and H b denote the width and height of the bounding box, and W and
H are the width and height of the image, respectively.

Line Intersection
In the light of the proposed painting method, it is desired to have more interac-
tion between agents. The interaction level of agents is reflected on the number of
line intersections in the image. Accordingly, we define the second part of fitness
evaluation by

f intersect = #intersections
WH

.

The fitness value of a painting is defined by

F = f edge + f intersect .