Image-Guided Rendering with an Evolutionary Algorithm Based on Cloud Model

Algorithm 1

The iREC algorithm.

Input: The original image and the parameters.

Output: The rendering image and animation.

(1) Set initialization parameters, the current image , the flag buffer , and .

(2) Pre-process the original image, and obtain (i phase, see Section 3.2).

(3) Generate an initial population () according to the forward normal cloud generator.

(4) repeat

(5) for    do

(6) Calculate the fitness value of each individual in the th population.

(7) Determine the top elite individuals according to the fitness values of individuals.

(8) for    do

(9) Draw with the th individual of the th population according to Algorithm 2 (R phase, see Section 3.3).

(10)end for

(11)Fitness-based reproduction (E phase, see Section 3.5), including selection, crossover, mutation

models using backward normal cloud generator.

(12)end for

(13)Produce the current image as a frame of rendering animation and then .

(14)Generate the new population () according to the forward normal cloud generator (C phase, see Section 3.4).

(15) until Best individual is good enough or stopping criterion is satisfied.

(16) Produce the current image as the final rendering image.