main.py
10.9 KB
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
206
207
208
209
210
211
212
213
214
215
216
217
218
219
220
221
222
223
224
225
226
227
228
229
230
231
232
233
234
235
236
237
238
239
240
241
242
243
244
245
246
247
248
249
250
251
252
253
254
255
256
257
258
259
260
261
262
263
264
265
import os
import cv2
import time
import json
import base64
import numpy as np
import data_util
import load_util
from src import face_detector
from src import face_landmark
from src import direction_classifier
from src import face_id
from src import doc_det
from src import abnormal_face
class InitModel(object):
def __init__(self, config):
face_det_model_path = config['FACE_DET']['MODEL_PATH']
face_landmark_model_path = config['FACE_LANDMARK_DET']['MODEL_PATH']
face_id_model_path = config['FACE_RECOGNIZE']['MODEL_PATH']
id_rotation_model_path = config['ID_DIRECTION']['MODEL_PATH']
doc_det_model_path = config['DOC_DET']['MODEL_PATH']
abnormal_face_model_path = config['ABNORMAL_FACE']['MODEL_PATH']
self.face_det_thr = config['FACE_DET']['FACE_THR']
self.is_rot_id = config['SERVICE']['ROT_ID']
self.is_rot_ids = config['SERVICE']['ROT_IDS']
self.is_det_id = config['SERVICE']['DET_ID']
self.is_det_ids = config['SERVICE']['DET_IDS']
self.return_location = config['SERVICE']['RETURN_LOCATION']
self.doc_det = config['SERVICE']['DOC_DET']
self.abnormal_face = config['SERVICE']['ABNORMAL_FACE']
self.direction_classifier = direction_classifier.Direction_Classifier(id_rotation_model_path)
self.face_recognizer = face_id.Face_Recognizer(face_id_model_path)
self.face_detector = face_detector.Face_Detector(face_det_model_path)
self.face_landmark_detector = face_landmark.Landmark_Detector(face_landmark_model_path)
self.doc_detector = doc_det.Doc_Detector(doc_det_model_path)
self.abnormal_recognizer = abnormal_face.Abnormal_Face(abnormal_face_model_path)
def error_return(self, result_dict, code):
result_dict['match'] = False
result_dict['face_sim'] = 0
result_dict['more'] = False
result_dict['code'] = code
result_dict['sim'] = 0
return result_dict
def recognition(self, id_rgb_image, life_rgb_image, config, compare_type):
result_dict = {'code':2000}
face_bboxes = []
face_scores = []
faces = []
if compare_type == 0:
thr = config['FACE_RECOGNIZE']['ID_ID_THR']
elif compare_type == 1:
thr = config['FACE_RECOGNIZE']['ID_LIFE_THR']
elif compare_type == 2:
thr = config['FACE_RECOGNIZE']['LIFE_LIFE_THR']
if self.doc_det:
if compare_type == 0:
id_doc_boxes, id_doc_classes, id_doc_scores = self.doc_detector.detect(id_rgb_image)
if len(id_doc_scores) == 0:
return self.error_return(result_dict, 4001)
id_doc_max = max(id_doc_scores)
id_doc_max_idx = id_doc_scores.tolist().index(id_doc_max)
id_doc_max_box = id_doc_boxes[id_doc_max_idx]
id_rgb_image = id_rgb_image[id_doc_max_box[1]:id_doc_max_box[3], id_doc_max_box[0]:id_doc_max_box[2],]
life_doc_boxes, life_doc_classes, life_doc_scores = self.doc_detector.detect(life_rgb_image)
if len(life_doc_scores) == 0:
return self.error_return(result_dict, 4002)
life_doc_max = max(life_doc_scores)
life_doc_max_idx = life_doc_scores.tolist().index(life_doc_max)
life_doc_max_box = life_doc_boxes[life_doc_max_idx]
life_rgb_image = life_rgb_image[life_doc_max_box[1]:life_doc_max_box[3], life_doc_max_box[0]:life_doc_max_box[2],]
result_dict['id_doc_bboxes'] = id_doc_boxes.astype(np.int32).tolist()
result_dict['life_doc_bboxes'] = life_doc_boxes.astype(np.int32).tolist()
elif compare_type == 1:
id_doc_boxes, id_doc_classes, id_doc_scores = self.doc_detector.detect(id_rgb_image)
if len(id_doc_scores) == 0:
return self.error_return(result_dict, 4001)
id_doc_max = max(id_doc_scores)
id_doc_max_idx = id_doc_scores.tolist().index(id_doc_max)
id_doc_max_box = id_doc_boxes[id_doc_max_idx]
id_rgb_image = id_rgb_image[id_doc_max_box[1]:id_doc_max_box[3], id_doc_max_box[0]:id_doc_max_box[2],]
result_dict['id_doc_bboxes'] = id_doc_boxes.astype(np.int32).tolist()
id_h, id_w, id_c = id_rgb_image.shape
if compare_type == 0 and self.is_rot_id:
id_rgb_image, id_direction_index = self.direction_classifier.reg(id_rgb_image)
life_rgb_image, life_direction_index = self.direction_classifier.reg(life_rgb_image)
result_dict['id_direction'] = int(id_direction_index)
result_dict['life_direction'] = int(life_direction_index)
elif compare_type == 1 and self.is_rot_id:
id_rgb_image, id_direction_index = self.direction_classifier.reg(id_rgb_image)
result_dict['id_direction'] = int(id_direction_index)
id_h, id_w, id_c = id_rgb_image.shape
life_h, life_w, life_c = life_rgb_image.shape
id_face_bboxes, id_face_landmarks, id_max_idx = self.face_detector.detect(id_rgb_image, self.face_det_thr)
life_face_bboxes, life_face_landmarks, life_max_idx = self.face_detector.detect(life_rgb_image, self.face_det_thr)
if compare_type != 0 and self.abnormal_face:
new_life_face_bboxes = []
print('1'*100)
for box_idx, face_bbox in enumerate(life_face_bboxes):
abnormal_pre = self.abnormal_recognizer.reg(life_rgb_image, face_bbox)
if abnormal_pre == 1:
new_life_face_bboxes.append(face_bbox)
life_face_bboxes = new_life_face_bboxes
if not id_face_bboxes:
return self.error_return(result_dict, 4003)
if not life_face_bboxes:
return self.error_return(result_dict, 4004)
result_dict['id_face_bboxes'] = id_face_bboxes
result_dict['id_face_landmarks'] = id_face_landmarks
result_dict['life_face_bboxes'] = life_face_bboxes
result_dict['life_face_landmarks'] = life_face_landmarks
id_face_landmark, id_face = self.face_landmark_detector.detect(id_rgb_image, id_face_bboxes[id_max_idx])
id_face_landmark = [id_face_landmark[104], id_face_landmark[105], id_face_landmark[46], id_face_landmark[84], id_face_landmark[90]]
id_norm_image = data_util.get_norm_face(id_face, id_face_landmark)
norm_images = [id_norm_image]
for f, life_face_bbox in enumerate(life_face_bboxes):
life_face_landmark, life_face = self.face_landmark_detector.detect(life_rgb_image, life_face_bbox)
life_face_landmark = [life_face_landmark[104], life_face_landmark[105], life_face_landmark[46], life_face_landmark[84], life_face_landmark[90]]
life_norm_image = data_util.get_norm_face(life_face, life_face_landmark)
norm_images.append(life_norm_image)
embeddings = self.face_recognizer.recognize(norm_images)
gallery_vector = np.mat(embeddings[0])
res = False
sim = 0
for p in range(1, len(embeddings)):
compare_vector = np.mat(embeddings[p])
dot = np.sum(np.multiply(gallery_vector, compare_vector), axis=1)
norm = np.linalg.norm(gallery_vector, axis=1) * np.linalg.norm(compare_vector, axis=1)
dist_1 = dot / norm
sim = dist_1.tolist()
sim = sim[0][0]
if sim > thr:
res = True
result_dict['id_index'] = id_max_idx
result_dict['life_index'] = p-1
result_dict['sim'] = sim
print('sim {} : {}'.format(p, sim))
result_dict['match'] = res
return result_dict
def main(id_base64_image, life_base64_image, compare_type):
if format_type == 0:
life_rgb_image = data_util.base64_to_img(life_base64_image)
id_rgb_image = data_util.base64_to_img(id_base64_image)
result_dict = inited_model.recognition(id_rgb_image, life_rgb_image, config, compare_type)
elif format_type == 1:
result_dict = inited_model.recognition(id_base64_image, life_base64_image, config, compare_type)
with open("tmp/test.json", "w") as f:
json.dump(result_dict, f)
return result_dict
if __name__ == '__main__':
image_dir = r'/data2/face_id/situ_other/pipeline_test/'
image_list_txt_path = r'/data2/face_id/situ_other/test2.txt'
config_path = r'config.yaml'
config = load_util.load_config(config_path)
inited_model = InitModel(config)
compare_type = config['SERVICE']['SCHEMA'] # 0:id-id 1: id-life 2: life-life
format_type = config['SERVICE']['FORMAT'] # 0:base64 1: RGB 2: URL
image_list_txt = open(image_list_txt_path, 'r')
image_list_txt_lines = image_list_txt.readlines()
hit = 0.01
hit_pos = 0.01
pre_pos = 0.01
pre_all = 0.01
positive_num = 0.01
for image_list_txt_line in image_list_txt_lines:
arr = image_list_txt_line.strip().split(' ')
label = arr[-1]
pre_all += 1
if label == '1':
positive_num += 1
id_image_name_arr = arr[0].split('/')
id_image_name = id_image_name_arr[-1]
id_image_name = id_image_name[2:]
id_image_dir = id_image_name_arr[-2]
id_image_path = os.path.join(image_dir, id_image_dir, id_image_name)
life_image_name_arr = arr[1].split('/')
life_image_name = life_image_name_arr[-1]
life_image_name = life_image_name[2:]
life_image_dir = life_image_name_arr[-2]
life_image_path = os.path.join(image_dir, life_image_dir, life_image_name)
id_image = cv2.imread(id_image_path)
life_image = cv2.imread(life_image_path)
r, id_image_str = cv2.imencode('.jpg', id_image)
id_base64_image = base64.b64encode(id_image_str)
r, life_image_str = cv2.imencode('.jpg', life_image)
life_base64_image = base64.b64encode(life_image_str)
st = time.time()
result_dict = {}
if format_type == 0:
result_dict = main(id_base64_image, life_base64_image, compare_type)
elif format_type == 1:
result_dict = main(id_image, life_image, compare_type)
print(result_dict)
et = time.time()
print('total time cost:{}'.format(round((et-st), 2)))
print(label)
res = result_dict['match']
if res:
if label == '1':
hit_pos += 1
hit += 1
pre_pos += 1
else:
if label == '0':
hit += 1
print('precision:{}'.format(hit_pos/pre_pos))
print('recall:{}'.format(hit_pos/positive_num))
print('accuracy:{}'.format(hit/pre_all))