-
Notifications
You must be signed in to change notification settings - Fork 0
/
genetic_impl.py
205 lines (170 loc) · 6.84 KB
/
genetic_impl.py
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
138
139
140
141
142
143
144
145
146
147
148
149
150
151
152
153
154
155
156
157
158
159
160
161
162
163
164
165
166
167
168
169
170
171
172
173
174
175
176
177
178
179
180
181
182
183
184
185
186
187
188
189
190
191
192
193
194
195
196
197
198
199
200
201
202
203
204
205
import random
import cv2
import numpy as np
import argparse
import os
# p = 5000, 500, 50
# m = 0.0001, 0.001
population_size = 500
max_iter = 200 # number of generations
mutation_rate = 0.0001
move = [[0, 0], [0, 1], [-1, 1], [-1, 0], [-1, -1], [0, -1], [1, -1], [1, 0], [1, 1]]
cost_values = [0, 45, 90, 135, 180]
best_score, best_result = 0.0, None # best score and solution of all generations
class Individual:
"""
chromosome: 1D array for each individual
fitness_score: float number for each individual's solution
"""
chromosome = []
fitness_score = 0
def __init__(self, dimension, size):
"""
initialize individuals
:param dimension: given size of environment exp: 20
:param size: individual's size exp: 20x20
condition: matrix of solution with each individual moves
total_cost: cost of individual moves
"""
self.chromosome = np.random.randint(len(move), size=size)
self.condition = np.full((dimension, dimension), 255, dtype=int)
self.total_cost = 0
prev = [0, 0]
start = [dimension - 1, 0]
self.condition[start[0], start[1]] = 0
for i in self.chromosome:
act = [x + y for x, y in zip(move[i], start)]
if not (act[0] < 0 or act[1] < 0 or act[0] > dim - 1 or act[1] > dim - 1):
self.total_cost += self.calculate_cost(move[i], prev)
start = act
prev = move[i]
self.condition[act[0], act[1]] = 0
self.condition = self.condition.reshape(dim * dim)
def fitness(self):
"""
calculate how individual is close a given solution via comparison of matrices
original image and individual's solutions are compared
:return: individual's fitness score
"""
score = 0
for i in range(len(target)):
if self.condition[i] == target[i]:
score += 1
self.fitness_score = score / len(target)
def crossing_over(self, parent, dimension):
"""
crossing over for desired individuals
:param parent: chromosome array to exchange genes
:param dimension: size of environment to create individual
:return:
"""
child = Individual(dimension, size)
rand = random.randrange(0, len(target))
for i in range(len(target)):
if i > rand:
child.chromosome[i] = self.chromosome[i]
else:
child.chromosome[i] = parent.chromosome[i]
return child
def mutation(self):
"""
change each individual's gene with random index and value
:return:
"""
for i in range(len(target)):
if random.random() <= mutation_rate:
# self.chromosome[i] = random.choice([0, 255])
self.chromosome[i] = np.random.randint(len(move), size=1)
@staticmethod
def save_best(filename, img, p, m, s):
"""
save the best solution as image file into related folder
:param filename: given file as original image (target)
:param img: best solution or individual to given file
:param p: population size
:param m: mutation rate
:param s: score
:return:
"""
path = f'results/{filename.split(".")[0]}'
os.makedirs(path, exist_ok=True)
cv2.imwrite(f'{path}/{p}_{m}_{s}_{filename}', img)
@staticmethod
def calculate_cost(next_m, first_m):
"""
calculate cost of individual
:param next_m: individual's next step that generated as random
:param first_m: individual's first step that generated as random
multiply difference value as index in cost values array
:return:
"""
diff = []
zip_object = zip(next_m, first_m)
for i, j in zip_object:
diff.append(abs(i - j))
if (diff[0] == 0 and diff[1] != 0) or (diff[0] != 0 and diff[1] == 0):
if first_m == [0, 0]:
return cost_values[0]
return cost_values[4]
return cost_values[sum(diff)]
if __name__ == '__main__':
parser = argparse.ArgumentParser()
parser.add_argument("-f", "--file", default='deneme.jpg', help='image name')
parser.add_argument("-d", "--dimension", default=20, help='environment size')
args = vars(parser.parse_args())
file = args['file'] # get filename
dim = int(args['dimension']) # get size of environment
"""
read image via opencv imread function, convert image to binary format and
show until generations end
"""
target = cv2.imread(f"images/{file}", 0).reshape(dim * dim)
(thresh, target) = cv2.threshold(target, 128, 255, cv2.THRESH_BINARY | cv2.THRESH_OTSU)
cv2.imshow("Original image", target.reshape(dim, dim))
size = dim * dim # individual's size
population = [Individual(dim, size) for i in range(population_size)] # create population
iterations = 0
while True:
iterations += 1
# find all individual's fitness score
for i in population:
i.fitness()
local_best = 0 # each generation's bests solution
for i in population:
if local_best < i.fitness_score:
local_best = i.fitness_score
if best_score == 0.0:
best_score = local_best
# if generated image greater than best, set best score as image
generated_image = np.array(i.condition.reshape(dim, dim), dtype=np.uint8)
if best_score < local_best:
best_score = local_best
best_result = generated_image
cv2.imshow("Generated image", generated_image) # show solution
cv2.waitKey(1)
if iterations == max_iter:
cv2.waitKey(1)
break
print("Generation:", iterations + 1, "\tScore:", str(round(local_best, 4)))
print("Current best score:", best_score)
new_population = []
"""
the higher the fitness function of an individual, the more it takes place in the new population
"""
for i in range(len(population)):
n = int(population[i].fitness_score * 100) + 1
for j in range(n):
new_population.append(population[i])
"""
crossing over and mutations are done here
"""
for i in range(len(population)):
i1 = random.randrange(0, len(new_population))
i2 = random.randrange(0, len(new_population))
new_i1 = new_population[i1]
new_i2 = new_population[i2]
child = new_i1.crossing_over(new_i2, dim)
child.mutation()
population[i] = child
# save global best image (max score) for all generations
Individual.save_best(filename=file, img=best_result, p=population_size, m=mutation_rate, s=best_score)