-
Notifications
You must be signed in to change notification settings - Fork 0
/
transformations.py
205 lines (158 loc) · 6.6 KB
/
transformations.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 queue
import threading
import time
from abc import ABC, abstractmethod, abstractproperty
import cv2
import numpy as np
from PyQt5 import QtWidgets, QtCore
import edit_grid
from utils import map_value
class TransformationPicker:
def __init__(self, edit_grid, parent, name):
self.edit_grid = edit_grid
self.dropdown_container = QtWidgets.QHBoxLayout()
dropdown_l = QtWidgets.QLabel(name)
self.dropdown = QtWidgets.QComboBox()
self.dropdown.addItems([cls.__name__ for cls in Transformer.__subclasses__()])
self.dropdown.currentTextChanged.connect(self.update_transformation)
self.dropdown_container.addWidget(dropdown_l)
self.dropdown_container.addWidget(self.dropdown)
parent.addLayout(self.dropdown_container)
self.slider_container = QtWidgets.QHBoxLayout()
slider_label = QtWidgets.QLabel('strength')
self.slider = QtWidgets.QSlider(QtCore.Qt.Horizontal)
self.slider.setRange(0, 100)
self.slider.valueChanged[int].connect(self.slider_update)
self.slider_container.addWidget(slider_label)
self.slider_container.addWidget(self.slider)
parent.addLayout(self.slider_container)
self.transformer = None
self.update_transformation()
def update_transformation(self):
cls = globals()[self.dropdown.currentText()]
self.transformer = cls()
self.slider.setValue(self.transformer.get_strength())
def slider_update(self, value):
self.transformer.set_strength(value)
self.edit_grid.update()
class Transformer(ABC):
@abstractmethod
def get_strength(self) -> float:
pass
@abstractmethod
def set_strength(self, value: float) -> float:
pass
@abstractmethod
def transform_image(self, image: np.ndarray, location: float) -> None:
pass
class TransformerScale(Transformer):
def __init__(self):
self.strength = 50
def get_strength(self) -> float:
return self.strength
def set_strength(self, value: float) -> None:
self.strength = value
def transform_image(self, image: np.ndarray, location: float) -> np.ndarray:
scale = (map_value(self.strength, 0, 100, 0.5, 1.5) + 1) * map_value(location, -1, 1, 0, 1)
if scale == 0:
return image.copy()
out = image.copy()
start_shape = out.shape
dim = (int(out.shape[1] * scale), int(out.shape[0] * scale))
if scale > 1:
out = cv2.resize(out, dim, interpolation=cv2.INTER_AREA)
x_crop = (dim[0] - start_shape[1]) // 2
y_crop = (dim[1] - start_shape[0]) // 2
out = out[y_crop:dim[1] - y_crop, x_crop:dim[0] - x_crop]
else:
out = cv2.resize(out, dim, interpolation=cv2.INTER_AREA)
pad_x = (start_shape[1] - dim[0]) // 2
pad_y = (start_shape[0] - dim[1]) // 2
in_img_h, in_img_w, channels = out.shape
out_img_h = in_img_h + (pad_y * 2)
out_img_w = in_img_w + (pad_x * 2)
mask = np.zeros((out_img_h, out_img_w, 1), np.uint8)
fill_color = (255, 255, 255)
out_img = cv2.copyMakeBorder(out, pad_y, pad_y, pad_x, pad_x, cv2.BORDER_CONSTANT, value=fill_color)
mask = cv2.rectangle(mask, (0, 0), (pad_x - 1, out_img_h - 1), (255), -1)
mask = cv2.rectangle(mask, (0, 0), (out_img_w - 1, pad_y - 1), (255), -1)
mask = cv2.rectangle(mask, (0, out_img_h - pad_y), (out_img_w - 1, out_img_h - 1), (255), -1)
mask = cv2.rectangle(mask, (out_img_w - pad_x, 0), (out_img_w - 1, out_img_h - 1), (255), -1)
out = cv2.inpaint(out_img, mask, 3, cv2.INPAINT_TELEA)
if out.shape[1] != 1024 or out.shape[0] != 1024:
out = cv2.resize(out, (1024, 1024), interpolation=cv2.INTER_AREA)
return out
class ComputeImageData:
def __init__(
self,
image: np.ndarray,
locations: [(float, float)],
transformer_x: Transformer,
transformer_y: Transformer,
thread_count: int = 1
):
self.src = image
self.locations = locations
self.transformer_x = transformer_x
self.transformer_y = transformer_y
self.thread_count = thread_count
self.output_images = []
self.exit_flag = False
self.lock = threading.Lock()
self.lock.acquire()
self.index_queue = queue.Queue()
for i in range(len(self.locations)):
self.index_queue.put(i)
# print(list(self.index_queue))
self.indexes = []
for i in range(self.thread_count):
self.indexes.append(-1)
self.lock.release()
def compute_image(thread_id, data: ComputeImageData) -> None:
while not data.exit_flag:
data.indexes[thread_id] = -1
data.lock.acquire()
if not data.index_queue.empty():
data.indexes[thread_id] = data.index_queue.get()
# print(thread_name + ' working on index ' + str(index))
data.lock.release()
if data.indexes[thread_id] < 0 or data.indexes[thread_id] >= len(data.locations):
break
else:
image = data.src
image = data.transformer_x.transform_image(image, data.locations[data.indexes[thread_id]][1])
image = data.transformer_y.transform_image(image, data.locations[data.indexes[thread_id]][0])
if image is not None:
data.lock.acquire()
data.output_images.append((data.indexes[thread_id], image))
data.lock.release()
# print(thread_name + ' done with index ' + str(index))
else:
print("Warning! Something went wrong!")
def compute_images(data: ComputeImageData) -> [np.ndarray]:
threads = []
for i in range(data.thread_count):
threads.append(ComputeImageThread(i, 'compute-image-thread-' + str(i), data))
threads[-1].start()
while not data.index_queue.empty():
pass
timeout_count = 1000
while len(data.output_images) != len(data.locations) and timeout_count > 0:
time.sleep(1)
timeout_count -= 1
data.exit_flag = True
out = []
for i in range(len(data.output_images)):
# print(data.output_images[i])
out.append(None)
for a in data.output_images:
out[a[0]] = a[1]
return out
class ComputeImageThread(threading.Thread):
def __init__(self, thread_id, name, data):
threading.Thread.__init__(self)
self.thread_id = thread_id
self.name = name
self.data = data
def run(self):
compute_image(self.thread_id, self.data)