-
Notifications
You must be signed in to change notification settings - Fork 114
/
yoloface_detect_align_module.py
181 lines (163 loc) · 7.43 KB
/
yoloface_detect_align_module.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
from yoloface.nn.mobilenetv3 import mobilenetv3_large, mobilenetv3_large_full, mobilenetv3_small
from yoloface.nn.models import DarknetWithShh
import torch
from yoloface.utils import non_max_suppression
import cv2
import numpy as np
from yoloface.hyp import hyp
from align_faces import align_process
class yoloface():
def __init__(self, net_type = 'mbv3_small_1_light', device = 'cuda', align=False):
self.long_side = -1 # -1 mean origin shape
backone = None
assert net_type in ['mbv3_small_1', 'mbv3_small_75', 'mbv3_large_1', 'mbv3_large_75',
"mbv3_large_75_light", "mbv3_large_1_light", 'mbv3_small_75_light', 'mbv3_small_1_light',
]
if net_type.startswith("mbv3_small_1"):
backone = mobilenetv3_small()
elif net_type.startswith("mbv3_small_75"):
backone = mobilenetv3_small(width_mult=0.75)
elif net_type.startswith("mbv3_large_1"):
backone = mobilenetv3_large()
elif net_type.startswith("mbv3_large_75"):
backone = mobilenetv3_large(width_mult=0.75)
elif net_type.startswith("mbv3_large_f"):
backone = mobilenetv3_large_full()
if 'light' in net_type:
net = DarknetWithShh(backone, hyp, light_head=True).to(device)
else:
net = DarknetWithShh(backone, hyp).to(device)
self.point_num = hyp['point_num']
weights = "yoloface/weights/{}_final.pt".format(net_type)
net.load_state_dict(torch.load(weights, map_location=device)['model'])
self.net = net.eval()
self.align = align
self.device = device
def detect(self, srcimg):
ori_h, ori_w, _ = srcimg.shape
LONG_SIDE = self.long_side
if self.long_side == -1:
max_size = max(ori_w, ori_h)
LONG_SIDE = max(32, max_size - max_size % 32)
if ori_h > ori_w:
scale_h = LONG_SIDE / ori_h
tar_w = int(ori_w * scale_h)
tar_w = tar_w - tar_w % 32
tar_w = max(32, tar_w)
tar_h = LONG_SIDE
else:
scale_w = LONG_SIDE / ori_w
tar_h = int(ori_h * scale_w)
tar_h = tar_h - tar_h % 32
tar_h = max(32, tar_h)
tar_w = LONG_SIDE
scale_w = tar_w * 1.0 / ori_w
scale_h = tar_h * 1.0 / ori_h
image = cv2.resize(srcimg, (tar_w, tar_h))
image = image[..., ::-1]
image = image.astype(np.float64)
# image = (image - hyp['mean']) / hyp['std']
image /= 255.0
image = np.transpose(image, [2, 0, 1])
image = np.expand_dims(image, axis=0)
with torch.no_grad():
image = torch.from_numpy(image)
image = image.to(self.device).float()
pred = self.net(image)[0]
pred = non_max_suppression(pred, 0.3, 0.35, multi_label=False, classes=0, agnostic=False, land=True,
point_num=self.point_num)
try:
det = pred[0].cpu().detach().numpy()
srcimg = srcimg.astype(np.uint8)
det[:, :4] = det[:, :4] / np.array([scale_w, scale_h] * 2)
det[:, 5:5 + self.point_num * 2] = det[:, 5:5 + self.point_num * 2] / np.array([scale_w, scale_h] * self.point_num)
except:
det = []
drawimg, face_rois = srcimg.copy(), []
for b in det:
# text = "{:.4f}".format(b[4])
b = list(map(int, b)) ###landmarks: numpy array, n x 10 (x1, y1 ... x5,y5)
cv2.rectangle(drawimg, (b[0], b[1]), (b[2], b[3]), (0, 0, 255), thickness=2)
# cx, cy = b[0], b[1] + 12
# cv2.putText(drawimg, text, (cx, cy), cv2.FONT_HERSHEY_DUPLEX, 0.5, (255, 255, 255))
cv2.circle(drawimg, (b[5], b[6]), 1, (0, 0, 255), 4)
cv2.circle(drawimg, (b[7], b[8]), 1, (0, 255, 255), 4)
cv2.circle(drawimg, (b[9], b[10]), 1, (255, 0, 255), 4)
cv2.circle(drawimg, (b[11], b[12]), 1, (0, 255, 0), 4)
cv2.circle(drawimg, (b[13], b[14]), 1, (255, 0, 0), 4)
# for i in range(5):
# cv2.putText(drawimg, str(i), (b[2*i+5], b[2*i+6]+12), cv2.FONT_HERSHEY_DUPLEX, 1, (0, 0, 255))
face_roi = srcimg[b[1]:b[3], b[0]:b[2]]
if self.align:
face_roi = align_process(srcimg, np.array(b[:4]), np.array(b[5:15]).reshape(-1, 2), (224,224))
face_rois.append(face_roi)
return drawimg, face_rois
def get_face(self, srcimg):
ori_h, ori_w, _ = srcimg.shape
LONG_SIDE = self.long_side
if self.long_side == -1:
max_size = max(ori_w, ori_h)
LONG_SIDE = max(32, max_size - max_size % 32)
if ori_h > ori_w:
scale_h = LONG_SIDE / ori_h
tar_w = int(ori_w * scale_h)
tar_w = tar_w - tar_w % 32
tar_w = max(32, tar_w)
tar_h = LONG_SIDE
else:
scale_w = LONG_SIDE / ori_w
tar_h = int(ori_h * scale_w)
tar_h = tar_h - tar_h % 32
tar_h = max(32, tar_h)
tar_w = LONG_SIDE
scale_w = tar_w * 1.0 / ori_w
scale_h = tar_h * 1.0 / ori_h
image = cv2.resize(srcimg, (tar_w, tar_h))
image = image[..., ::-1]
image = image.astype(np.float64)
# image = (image - hyp['mean']) / hyp['std']
image /= 255.0
image = np.transpose(image, [2, 0, 1])
image = np.expand_dims(image, axis=0)
with torch.no_grad():
image = torch.from_numpy(image)
image = image.to(self.device).float()
pred = self.net(image)[0]
pred = non_max_suppression(pred, 0.3, 0.35, multi_label=False, classes=0, agnostic=False, land=True,
point_num=self.point_num)
try:
det = pred[0].cpu().detach().numpy()
srcimg = srcimg.astype(np.uint8)
det[:, :4] = det[:, :4] / np.array([scale_w, scale_h] * 2)
det[:, 5:5 + self.point_num * 2] = det[:, 5:5 + self.point_num * 2] / np.array([scale_w, scale_h] * self.point_num)
except:
det = []
boxs, face_rois = [], []
for b in det:
b = list(map(int, b))
del b[4] ### delte score
boxs.append(b)
face_roi = srcimg[b[1]:b[3], b[0]:b[2]]
if self.align:
face_roi = align_process(srcimg, np.array(b[:4]), np.array(b[4:14]).reshape(-1, 2), (224, 224))
face_rois.append(face_roi)
return boxs, face_rois
if __name__ == "__main__":
device = 'cuda' if torch.cuda.is_available() else 'cpu'
yoloface_detect = yoloface(device=device, align=True)
imgpath = 's_l.jpg'
srcimg = cv2.imread(imgpath)
drawimg, face_rois = yoloface_detect.detect(srcimg)
# boxs, face_rois = yoloface_detect.get_face(srcimg)
# drawimg = srcimg.copy()
# for i,box in enumerate(boxs):
# cv2.rectangle(drawimg, (box[0], box[1]), (box[2], box[3]), (0, 0, 255), thickness=2)
# for j in range(5):
# cv2.circle(drawimg, (box[4+j * 2], box[4+j * 2 + 1]), 2, (0, 255, 0), thickness=-1)
# for i,face in enumerate(face_rois):
# cv2.namedWindow('face'+str(i), cv2.WINDOW_NORMAL)
# cv2.imshow('face'+str(i), face)
cv2.namedWindow('detect', cv2.WINDOW_NORMAL)
cv2.imshow('detect', drawimg)
cv2.waitKey(0)
cv2.destroyAllWindows()