-
Notifications
You must be signed in to change notification settings - Fork 0
/
generator.py
168 lines (145 loc) · 6.47 KB
/
generator.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
import config
from PIL import ImageFont
from PIL import Image
from PIL import ImageDraw
import numpy as np
import cv2
import read_chinese
import argparse
parser = argparse.ArgumentParser()
parser.add_argument('--resolution', action="store", help='generate a specific resolution datasets')
parser.add_argument('--mode', action="store", help='generate mode: train or error-test')
def generate_doc_with_code_and_bias(bits: list,
text: str,
font_names: list,
bias={}) -> list:
"""Generate document with given text, bits to be encoded, and font names.
When text is longer than bits, font_names[0] will be used.
When text is shorter than bits, extra bits will be ignored.
Params:
bits: List of integers, e.g. [0, 1, 0, 0]. Don't have to be 0 or 1, because we can have more than two kinds of fonts.
text: e.g. "你好世界".
font_names: Name of font, or path of font file.
bias:
Returns:
List of PIL.Image.
"""
fonts = [ImageFont.truetype(f, config.FONT_SIZE) for f in font_names]
ret = []
row = col = 0
draw = None
tbias = [(0, 0) for _ in range(len(font_names))]
for k, v in bias.items():
tbias[k] = v
for cursor in range(len(text)):
# New page
if row == 0 and col == 0:
img = Image.new("RGB", config.DOC_IMAGE_SIZE, config.FOREGROUND_COLOR)
draw = ImageDraw.Draw(img)
draw.rectangle(xy=[
(config.BLANK_SIZE[0], config.BLANK_SIZE[1]),
(config.DOC_IMAGE_SIZE[0] - config.BLANK_SIZE[0],
config.DOC_IMAGE_SIZE[1] - config.BLANK_SIZE[1])
],
outline=config.BACKGROUND_COLOR,
width=config.BOARDER_SIZE)
ret.append(img)
if cursor >= len(bits):
bit = 0
else:
bit = bits[cursor]
draw.text((config.BLANK_SIZE[0] + config.BOARDER_SIZE + (col + 1) *
(config.FONT_SIZE + 2 * config.MARGIN_SIZE) +
config.MARGIN_SIZE + tbias[bit][0], config.BLANK_SIZE[1] + config.BOARDER_SIZE +
(row + 1) * (config.FONT_SIZE + 2 * config.MARGIN_SIZE) +
config.MARGIN_SIZE + tbias[bit][1]),
text[cursor],
config.BACKGROUND_COLOR,
font=fonts[bit])
tmp = col + 1
col = tmp % (config.DOC_DIM[0] - 2)
row = (row + tmp // (config.DOC_DIM[0] - 2)) % (config.DOC_DIM[1] - 2)
return ret
def generate_doc(chars: str, font_name: str):
print(font_name)
img = Image.new("RGB", config.DOC_IMAGE_SIZE, config.FOREGROUND_COLOR)
font = ImageFont.truetype(font_name, config.FONT_SIZE)
draw = ImageDraw.Draw(img)
draw.rectangle(xy=[(config.BLANK_SIZE[0], config.BLANK_SIZE[1]),
(config.DOC_IMAGE_SIZE[0] - config.BLANK_SIZE[0],
config.DOC_IMAGE_SIZE[1] - config.BLANK_SIZE[1])],
outline=config.BACKGROUND_COLOR,
width=config.BOARDER_SIZE)
try:
cursor = 0
for row in range(1, config.DOC_DIM[1] - 1):
for col in range(1, config.DOC_DIM[0] - 1):
draw.text((config.BLANK_SIZE[0] + config.BOARDER_SIZE + col *
(config.FONT_SIZE + 2 * config.MARGIN_SIZE) +
config.MARGIN_SIZE,
config.BLANK_SIZE[1] + config.BOARDER_SIZE + row *
(config.FONT_SIZE + 2 * config.MARGIN_SIZE) +
config.MARGIN_SIZE),
chars[cursor],
config.BACKGROUND_COLOR,
font=font)
cursor += 1
except IndexError:
pass
return img
def generate_iter(chars: str, font_name: str):
font_name = "data/fonts/{}.ttf".format(font_name)
for i in range(0, len(chars), config.DOC_DIM[0] * config.DOC_DIM[1]):
image = generate_doc(chars[i:], font_name)
# display(generate_doc(chars[i:], "fonts/{}".format(FONT_NAME), FONT_SIZE))
img = cv2.cvtColor(np.asarray(image), cv2.COLOR_RGB2BGR)
cv2.imshow("img", img)
cv2.waitKey()
def generate_doc_with_random_code(n: int):
bits = np.random.randint(2, size=(n, ))
print(bits)
text = read_chinese.read_chinese3000()
for i in range(len(text) // n):
text_epo = text[i * n: i * n + n]
# docs = generate_doc_with_code_and_bias(bits,
# text_epo, ['data/fonts/HuaWenSun.ttf', 'data/fonts/MicroSun.ttf'],
# bias={1: (0, 0)})
docs = generate_doc_with_code_and_bias(bits,
text_epo, ['data/fonts/HuaWenSun.ttf', 'data/fonts/MicroSun.ttf'],
bias={0: (0, -8)})
for img in docs:
img = cv2.cvtColor(np.asarray(img), cv2.COLOR_RGB2BGR)
cv2.imshow("img", img)
cv2.waitKey()
if __name__ == "__main__":
sep_img = Image.open("data/seperate.jpeg")
# 生成文档，敲回车生成下一页
chars = read_chinese.read_chinese3000()
chars = chars
# generate_doc(chars, config.FONT_NAME_Fangzheng)
NUM = len(chars)
code = [0 for _ in range(NUM)]
text = chars[:NUM]
args = parser.parse_args()
if args.mode == 'train':
# 微软仿宋
doc = generate_doc_with_code_and_bias(code,
text, ['data/fonts/MicroSun.ttf'],
bias={0: (0, 0)})
for img in doc:
img = cv2.cvtColor(np.asarray(img), cv2.COLOR_RGB2BGR)
cv2.imshow("img", img)
cv2.waitKey()
sep_img = cv2.cvtColor(np.asarray(sep_img), cv2.COLOR_RGB2BGR)
cv2.imshow("sep", sep_img)
cv2.waitKey()
# 华文仿宋
doc = generate_doc_with_code_and_bias(code,
text, ['data/fonts/HuaWenSun.ttf'],
bias={0: (0, -8)})
for img in doc:
img = cv2.cvtColor(np.asarray(img), cv2.COLOR_RGB2BGR)
cv2.imshow("img", img)
cv2.waitKey()
elif args.mode == 'error':
generate_doc_with_random_code((config.DOC_DIM[0] - 2) * (config.DOC_DIM[1] - 2))