-
Notifications
You must be signed in to change notification settings - Fork 0
/
dbface_detect_align_module.py
161 lines (144 loc) · 7.11 KB
/
dbface_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
import dbface.common as common
import torch
import numpy as np
import cv2
from dbface.DBFaceSmallH import DBFace as dbface_small
from dbface.DBFace import DBFace as dbface
import torch.nn.functional as F
from align_faces import align_process
class dbface_detect():
def __init__(self, net_type = 'dbface_small', device = 'cuda', align=False):
if net_type == 'dbface_small':
self.net = dbface_small().to(device)
self.net.load_state_dict(torch.load('dbface/dbfaceSmallH.pth', map_location=device))
self.net.eval()
else:
self.net = dbface().to(device)
self.net.load_state_dict(torch.load('dbface/dbface.pth', map_location=device))
self.net.eval()
self.align = align
self.device = device
self.mean = np.array([0.408, 0.447, 0.47], dtype=np.float32).reshape((1, 1, -1)) ###广播法则
self.std = np.array([0.289, 0.274, 0.278], dtype=np.float32).reshape((1, 1, -1))
def nms(self, objs, iou=0.5):
if objs is None or len(objs) <= 1:
return objs
objs = sorted(objs, key=lambda obj: obj.score, reverse=True)
keep = []
flags = [0] * len(objs)
for index, obj in enumerate(objs):
if flags[index] != 0:
continue
keep.append(obj)
for j in range(index + 1, len(objs)):
if flags[j] == 0 and obj.iou(objs[j]) > iou:
flags[j] = 1
return keep
def detect(self, srcimg, threshold=0.3, nms_iou=0.3):
image = common.pad(srcimg).astype(np.float32)
image = ((image / 255.0 - self.mean) / self.std)
image = image.transpose(2, 0, 1)
with torch.no_grad():
torch_image = torch.from_numpy(image)[None]
torch_image = torch_image.to(self.device)
hm, box, landmark = self.net(torch_image)
hm_pool = F.max_pool2d(hm, 3, 1, 1)
scores, indices = ((hm == hm_pool).float() * hm).view(1, -1).cpu().topk(1000)
hm_height, hm_width = hm.shape[2:]
scores = scores.squeeze()
indices = indices.squeeze()
ys = list((indices / hm_width).int().data.numpy())
xs = list((indices % hm_width).int().data.numpy())
scores = list(scores.data.numpy())
box = box.cpu().squeeze().data.numpy()
landmark = landmark.cpu().squeeze().data.numpy()
stride = 4
objs = []
for cx, cy, score in zip(xs, ys, scores):
if score < threshold:
break
x, y, r, b = box[:, cy, cx]
xyrb = (np.array([cx, cy, cx, cy]) + [-x, -y, r, b]) * stride
x5y5 = landmark[:, cy, cx]
x5y5 = (common.exp(x5y5 * 4) + ([cx] * 5 + [cy] * 5)) * stride
box_landmark = list(zip(x5y5[:5], x5y5[5:]))
objs.append(common.BBox(0, xyrb=xyrb, score=score, landmark=box_landmark))
objs = self.nms(objs, iou=nms_iou)
drawimg, face_rois = srcimg.copy(), []
for i, obj in enumerate(objs):
box, score, landmark = list(map(int, obj.box)), obj.score, obj.landmark
landmark = [int(x) for t in obj.landmark for x in t]
# landmark = sum(obj.landmark, ())
# landmark = list(itertools.chain.from_iterable(obj.landmark))
cv2.rectangle(drawimg, (box[0], box[1]), (box[2], box[3]), (0, 0, 255), thickness=2)
for j in range(0, 5):
cv2.circle(drawimg, (landmark[j * 2], landmark[j * 2 + 1]), 2, (0, 255, 0), thickness=-1)
# cv2.putText(drawimg, str(j), (landmark[j * 2], landmark[j * 2 + 1] + 12), cv2.FONT_HERSHEY_DUPLEX, 1, (0, 0, 255))
face_roi = srcimg[box[1]:box[3], box[0]:box[2]]
# import imutils
# angle = np.rad2deg(np.arctan2(landmark[3] - landmark[1], landmark[2] - landmark[0]))
# if angle != 0 and self.align:
# face_roi = imutils.rotate(face_roi, angle)
if self.align:
face_roi = align_process(srcimg, np.array(box), np.array(landmark).reshape(-1, 2), (224,224))
face_rois.append(face_roi)
return drawimg, face_rois
def get_face(self, srcimg, threshold=0.3, nms_iou=0.3):
image = common.pad(srcimg).astype(np.float32)
image = ((image / 255.0 - self.mean) / self.std)
image = image.transpose(2, 0, 1)
with torch.no_grad():
torch_image = torch.from_numpy(image)[None]
torch_image = torch_image.to(self.device)
hm, box, landmark = self.net(torch_image)
hm_pool = F.max_pool2d(hm, 3, 1, 1)
scores, indices = ((hm == hm_pool).float() * hm).view(1, -1).cpu().topk(1000)
hm_height, hm_width = hm.shape[2:]
scores = scores.squeeze()
indices = indices.squeeze()
ys = list((indices / hm_width).int().data.numpy())
xs = list((indices % hm_width).int().data.numpy())
scores = list(scores.data.numpy())
box = box.cpu().squeeze().data.numpy()
landmark = landmark.cpu().squeeze().data.numpy()
stride = 4
objs = []
for cx, cy, score in zip(xs, ys, scores):
if score < threshold:
break
x, y, r, b = box[:, cy, cx]
xyrb = (np.array([cx, cy, cx, cy]) + [-x, -y, r, b]) * stride
x5y5 = landmark[:, cy, cx]
x5y5 = (common.exp(x5y5 * 4) + ([cx] * 5 + [cy] * 5)) * stride
box_landmark = list(zip(x5y5[:5], x5y5[5:]))
objs.append(common.BBox(0, xyrb=xyrb, score=score, landmark=box_landmark))
objs = self.nms(objs, iou=nms_iou)
boxs, face_rois = [], []
for i, obj in enumerate(objs):
box, score, landmark = list(map(int, obj.box)), obj.score, obj.landmark
landmark = [int(x) for t in obj.landmark for x in t]
boxs.append(box+landmark)
face_roi = srcimg[box[1]:box[3], box[0]:box[2]]
if self.align:
face_roi = align_process(srcimg, np.array(box), np.array(landmark).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'
DBface_Detect = dbface_detect(device=device, align=True)
imgpath = 's_l.jpg'
srcimg = cv2.imread(imgpath)
drawimg, face_rois = DBface_Detect.detect(srcimg)
# boxs, face_rois = dbface_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()