-
Notifications
You must be signed in to change notification settings - Fork 2
/
visualization.py
148 lines (131 loc) · 6.3 KB
/
visualization.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
from autocrop import cropper
from autocrop.utils import resize_image_op
import cv2
import random
import functools
def draw_anchor_bboxes(img, bboxes, shifit=False, color=None, thickness=1):
change_color = True if color is None else False
for idx, bbox in enumerate(bboxes):
font = cv2.FONT_HERSHEY_TRIPLEX
if shifit:
dx1, dx2, dy1, dy2 = int((random.random() - 1) * 4), int((random.random() - 1) * 4), \
int((random.random() - 1) * 4), int((random.random() - 1) * 4)
else:
dx1, dx2, dy1, dy2 = [0, 0, 0, 0]
if change_color:
r, g, b = int(random.random() * 255), int(random.random() * 255), int(random.random() * 255)
color = (b, g, r)
if not isinstance(thickness, int):
thickness = int(thickness)
cv2.rectangle(img=img,
# xmin ymin xmax ymax
# xmin, ymin
# |
# |
# |___________xmax, ymax
pt1=tuple([bbox[0] + dx1, bbox[1] + dy1]),
pt2=tuple([bbox[2] + dx2, bbox[3] + dy2]),
color=color,
thickness=thickness)
cv2.putText(img, str(idx + 1),
tuple([bbox[0] + dx1, bbox[1] + dy1]),
font, 0.5, (0, 0, 0))
def bbox_visualization(img_path='imgs/demo.jpg',
crop_height=1,
crop_width=1):
"""
Visualize anchor bboxes:
:param img_path
:param rgb_img:
:param crop_height:
:param crop_width:
:return:
"""
auto_cropper = cropper.AutoCropper(model='mobilenetv2',
cuda=True,
use_face_detector=True)
img = cv2.imread(img_path)
input_img, scale_height, scale_width = resize_image_op(img)
face_bboxes = auto_cropper.detect_face(input_img)
raw_face_bboxes = []
for fbbox in face_bboxes:
tbbox = [int(round(scale_width * fbbox[0])),
int(round(scale_height * fbbox[1])),
int(round(scale_width * fbbox[2])),
int(round(scale_height * fbbox[3])),
]
raw_face_bboxes.append(tbbox)
print(raw_face_bboxes)
print([(x[2] - x[0]) * (x[3] - x[1]) for x in raw_face_bboxes])
"""
No filter
"""
img1 = img.copy()
generate_bbox_func_partial = functools.partial(auto_cropper.generate_anchor_bboxes,
image=input_img,
scale_height=scale_height,
scale_width=scale_width,
crop_height=crop_height,
crop_width=crop_width)
crop_func_partial = functools.partial(auto_cropper.crop,
rgb_image=img,
topK=1,
crop_width=crop_width,
crop_height=crop_height)
trans_bboxes, source_bboxes = generate_bbox_func_partial(face_bboxes=[],
single_face_center=False)
draw_anchor_bboxes(img=img1, bboxes=raw_face_bboxes, shifit=False, color=(0, 255, 0), thickness=3)
draw_anchor_bboxes(img=img1, bboxes=source_bboxes, shifit=True)
cv2.imshow('No Filter, {} Face, {} BBoxes'.format(len(face_bboxes),
len(source_bboxes)),
img1)
cv2.imwrite('No Filter, {} Face, {} BBoxes.jpg'.format(len(face_bboxes),
len(source_bboxes)),
img1)
ret1 = crop_func_partial(filter_face=False,
single_face_center=False)
bbox = ret1[0]
cv2.imshow("ret1", img[bbox[1]: bbox[3] + 1, bbox[0]: bbox[2] + 1, :])
cv2.imwrite("ret1.jpg", img[bbox[1]: bbox[3] + 1, bbox[0]: bbox[2] + 1, :])
"""
Filter Face
"""
img2 = img.copy()
trans_bboxes, source_bboxes = generate_bbox_func_partial(face_bboxes=face_bboxes,
single_face_center=False)
draw_anchor_bboxes(img=img2, bboxes=raw_face_bboxes, shifit=False, color=(0, 255, 0), thickness=3)
draw_anchor_bboxes(img=img2, bboxes=source_bboxes, shifit=True)
cv2.imshow('Filter Face, {} Face, {} BBoxes'.format(len(face_bboxes),
len(source_bboxes)),
img2)
cv2.imwrite('Filter Face, {} Face, {} BBoxes.jpg'.format(len(face_bboxes),
len(source_bboxes)),
img2)
ret2 = crop_func_partial(filter_face=True,
single_face_center=False)
bbox = ret2[0]
cv2.imshow("ret2", img[bbox[1]: bbox[3] + 1, bbox[0]: bbox[2] + 1, :])
cv2.imwrite("ret2.jpg", img[bbox[1]: bbox[3] + 1, bbox[0]: bbox[2] + 1, :])
"""
Filter Center Face
"""
img3 = img.copy()
trans_bboxes, source_bboxes = generate_bbox_func_partial(face_bboxes=face_bboxes,
single_face_center=True)
draw_anchor_bboxes(img=img3, bboxes=raw_face_bboxes, shifit=False, color=(0, 255, 0), thickness=3)
draw_anchor_bboxes(img=img3, bboxes=source_bboxes, shifit=True)
cv2.imshow('Filter Center Face, {} Face, {} BBoxes'.format(len(face_bboxes),
len(source_bboxes)),
img3)
cv2.imwrite('Filter Center Face, {} Face, {} BBoxes.jpg'.format(len(face_bboxes),
len(source_bboxes)),
img3)
ret3 = crop_func_partial(filter_face=True,
single_face_center=True)
bbox = ret3[0]
cv2.imshow("ret3", img[bbox[1]: bbox[3] + 1, bbox[0]: bbox[2] + 1, :])
cv2.imwrite("ret3.jpg", img[bbox[1]: bbox[3] + 1, bbox[0]: bbox[2] + 1, :])
cv2.waitKey()
cv2.destroyAllWindows()
if __name__ == '__main__':
bbox_visualization(img_path='imgs/demo.jpg', crop_height=478, crop_width=342)