-
Notifications
You must be signed in to change notification settings - Fork 0
/
utils.py
447 lines (371 loc) · 13.5 KB
/
utils.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
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
266
267
268
269
270
271
272
273
274
275
276
277
278
279
280
281
282
283
284
285
286
287
288
289
290
291
292
293
294
295
296
297
298
299
300
301
302
303
304
305
306
307
308
309
310
311
312
313
314
315
316
317
318
319
320
321
322
323
324
325
326
327
328
329
330
331
332
333
334
335
336
337
338
339
340
341
342
343
344
345
346
347
348
349
350
351
352
353
354
355
356
357
358
359
360
361
362
363
364
365
366
367
368
369
370
371
372
373
374
375
376
377
378
379
380
381
382
383
384
385
386
387
388
389
390
391
392
393
394
395
396
397
398
399
400
401
402
403
404
405
406
407
408
409
410
411
412
413
414
415
416
417
418
419
420
421
422
423
424
425
426
427
428
429
430
431
432
433
434
435
436
437
438
439
440
441
442
443
444
445
446
447
from __future__ import absolute_import, print_function
import os
import json
import numpy as np
import math
import cv2
from collections import defaultdict
import dlib
from imutils import face_utils
import imutils
import face_recognition
from six.moves import cPickle as pickle
from sklearn.neighbors import KDTree
import joblib
def format_table(stats):
table = [
" | T | F",
"-----------",
" P |{tp}|{fp}",
"-----------",
" N |{tn}|{fn}",
]
table = "\n".join(table)
vals = {}
for key, value in stats.items():
as_str = str(value)
if len(as_str) == 1:
as_str = " %s " % as_str
elif len(as_str) == 2:
as_str = " %s" % as_str
elif len(as_str) == 3:
as_str = "%s" % as_str
vals[key] = as_str
return table.format(**vals)
def format_graph(stats):
graph = [
"TP: {tp}",
"TN: {tn}",
"FP: {fp}",
"FN: {fn}",
]
graph = "\n".join(graph)
graph_stats = {}
for key, val in stats.items():
if len(key) == 2:
graph_stats[key] = "#"*val
return graph.format(**graph_stats)
class NoFaceException(Exception):
pass
class NoMatchException(Exception):
pass
MATCH_DATA_DIR = "match_data"
def get_people_list():
with open(os.path.join(MATCH_DATA_DIR, "people.json"), "rb") as f:
people = json.load(f)
return people
def write_people_list(people):
with open(os.path.join(MATCH_DATA_DIR, "people.json"), "wb") as f:
json.dump(people, f)
def read_match_data():
all_data = {}
for dir_name in os.listdir(MATCH_DATA_DIR):
full_dir = os.path.join(MATCH_DATA_DIR, dir_name)
if os.path.isdir(full_dir):
person_info = load_person(full_dir)
assert person_info['name'] not in all_data
all_data[person_info['name']] = person_info
return all_data
def load_person(directory):
file_names = os.listdir(directory)
assert 'info.json' in file_names
with open(os.path.join(directory, "info.json"), 'r') as f:
info = json.load(f)
info['encodings'] = []
for name in file_names:
if name.endswith(".png") or name.endswith(".jpg"):
encoding = load_and_cache_encoding(directory, name[:-4])
info['encodings'].append(encoding)
return info
def load_and_cache_encoding(directory, prefix, jitters=100):
bin_file = os.path.join(directory, prefix)
if os.path.isfile(bin_file + ".npy"):
encoding = np.load(bin_file + ".npy")
else:
if os.path.isfile(os.path.join(directory, prefix + ".png")):
encoding = compute_image_encoding(directory, prefix + ".png", jitters)
else:
encoding = compute_image_encoding(directory, prefix + ".jpg", jitters)
if encoding is None:
encoding = np.array(0)
np.save(bin_file, encoding)
if not encoding.any():
return None
return encoding
def compute_image_encoding(directory, filename, jitters=100):
"""
Assumes it will receive an image with a single face in it.
No more, no less.
"""
fullname = os.path.join(directory, filename)
img = face_recognition.load_image_file(fullname)
face_encodings = face_recognition.face_encodings(img, num_jitters=jitters)
if len(face_encodings) > 0:
return face_encodings[0]
else:
return None
def get_names_faces_lists(all_data):
names = []
encoded_faces = []
for name, person in all_data.items():
for ef in person['encodings']:
names.append(name)
encoded_faces.append(ef)
return names, encoded_faces
def get_identified_people(cv2_img, known_faces, names):
# Convert the image from BGR color (which OpenCV uses) to RGB color (which face_recognition uses)
rgb_frame = cv2_img[:, :, ::-1]
# Find all the faces and face enqcodings in the frame of video
face_locations = face_recognition.face_locations(rgb_frame)
face_encodings = face_recognition.face_encodings(rgb_frame, face_locations, num_jitters=10)
if len(face_encodings) == 0:
return {}
face_encoding = face_encodings[0]
matches = face_recognition.compare_faces(known_faces, face_encoding)
hits = defaultdict(int)
for match, name in zip(matches, names):
# FML that 'match' is a np bool, not 'is True'
if match == True:
hits[name] += 1
return hits
def get_encoding_from_cv2_img(cv2_img):
# Convert the image from BGR color (which OpenCV uses) to RGB color (which face_recognition uses)
rgb_frame = cv2_img[:, :, ::-1]
# Find all the faces and face encodings in the frame of video
face_locations = face_recognition.face_locations(rgb_frame)
face_encodings = face_recognition.face_encodings(rgb_frame, face_locations, num_jitters=10)
if len(face_encodings) == 0:
return None
return face_encodings[0]
def get_face_distances(cv2_img, known_faces):
face_encoding = get_encoding_from_cv2_img(cv2_img)
if face_encoding is not None:
return get_face_distances_with_encoding(face_encoding, known_faces)
else:
return None
def get_face_distances_with_encoding(face_encoding, known_faces):
#matches = face_recognition.compare_faces(known_faces, face_encoding)
distances = face_recognition.face_distance(known_faces, face_encoding)
return distances
def get_best_match_with_encoding(encoding, known_faces, names):
distances = get_face_distances_with_encoding(encoding, known_faces)
return _get_best_match(distances, names)
def get_best_match(cv2_img, known_faces, names):
distances = get_face_distances(cv2_img, known_faces)
return _get_best_match(distances, names)
TOLERANCE = 0.35
def _get_best_match(distances, names):
if distances is None:
raise NoFaceException()
#return None
min_dist = 1
use_name = None
for name, dist in zip(names, distances):
if dist < min_dist:
min_dist = dist
use_name = name
if min_dist > TOLERANCE:
raise NoMatchException()
#return None
else:
return use_name
def prompt_person(people_list):
names = set([x['name'] for x in people_list])
OFFSET = 3
print(" 1) Other")
print(" 2) Never match")
i = 0
i_name = {}
for i, person in enumerate(sorted(people_list, key=lambda x: x['name'])):
i_name[i+OFFSET] = person
print(" %s) %s" % (i+OFFSET, person['name']))
while True:
num = input("number? ")
try:
num = int(num)
if num > i+OFFSET or num < 1:
continue
except ValueError:
continue
break
if num == 1:
while True:
name = input("name? ")
email = input("email? ")
if name in names:
print("name already exists")
continue
if email == '':
email = name.lower()
if prompt_yn("%s - %s correct?" % (name, email)):
break
people_list.append({
'name': name,
'email': email+"@athinkingape.com",
})
write_people_list(people_list)
return name
elif num == 2:
return 'nevermatch'
else:
return i_name[num]['name']
def prompt_yn(text):
while True:
inp = input("%s ([y]/n)" % text)
if inp == '' or inp == 'y':
return True
elif inp == 'n':
return False
def prompt_list(list_data):
print("1) Other")
print("2) Nevermatch")
for i, row in enumerate(list_data):
print("%s) %s" % (i+3, row))
while True:
inp = input("Enter number: ")
try:
inp = int(inp)
except:
continue
print(inp)
if inp < 1 or inp > i+3:
continue
if inp == 2:
return "nevermatch"
elif inp == 1:
return None
else:
print(list_data[inp-3])
return list_data[inp-3]
def write_name(directory, dir_name, name):
label_path = os.path.join(directory, dir_name, "name.txt")
data = name+"\n"
with open(label_path, 'w') as f:
f.write(data)
def load_people(directory):
"""
returns a dictionary of name to list of face encodings
ignores all nevermatch image sets
"""
people = defaultdict(list)
for dir_name in os.listdir(directory):
fullpath = os.path.join(directory, dir_name)
if os.path.isdir(fullpath) and dir_name != "__pycache__":
names = os.listdir(fullpath)
if "name.txt" not in names:
continue
with open(os.path.join(fullpath, "name.txt"), "rb") as f:
truth_name = f.read().strip()
if truth_name == 'nevermatch':
continue
for img_name in names:
if img_name == 'name.txt':
continue
if img_name.endswith(".png") or img_name.endswith(".jpg"):
encoding = load_and_cache_encoding(fullpath, img_name[:-4], jitters=10)
if encoding is None:
continue
people[truth_name].append((encoding, os.path.join(fullpath, img_name)))
else:
continue
return people
class FaceNameStorage(object):
def __init__(self, names, faces, filepaths=None, name_email=None):
self.names = names
self.faces = faces
self.filepaths = filepaths
self.name_email = name_email
def get_info(self, index):
return {
'name': self.names[index],
'face': self.faces[index],
'filepath': self.filepaths[index],
}
class TreeModel(object):
MAX_DISTANCE = 0.27
NEIGHBOUR_COUNT = 10
MIN_NEIGHBOUR_COUNT = 5
def __init__(self, face_name_storage, neighbour_count=NEIGHBOUR_COUNT):
self.face_name_storage = face_name_storage
self.faces = np.asarray(self.face_name_storage.faces)
self.names = self.face_name_storage.names
self.tree = KDTree(self.faces)
self.neighbour_count = neighbour_count
def get_predictions(self, faces, max_distance=MAX_DISTANCE):
# query_radius doesn't like getting an empty array.
if len(faces) == 0:
return []
results = []
#distances_s, indices_s = self.tree.query(faces, k=self.neighbour_count)
indices_s = self.tree.query_radius(faces, r=max_distance, return_distance=False)
print(indices_s)
#for distances, indices in zip(distances_s, indices_s):
for indices in indices_s:
votes = defaultdict(int)
#for distance, index in zip(distances, indices):
for index in indices:
#if distance <= self.max_distance:
votes[self.names[index]] += 1
total_votes = sum(votes.values())
if total_votes < self.MIN_NEIGHBOUR_COUNT:
results.append(None)
continue
# If we didn't get at least 1/4 of our nearest neighbours inside our distance, no match
#if total_votes < self.NEIGHBOUR_COUNT/4:
#results.append(None)
#continue
best_person_name, best_person_votes = sorted(votes.items(), key=lambda x: x[1], reverse=True)[0]
# Make sure our vote leader has more than half of the total votes
if (best_person_votes / float(total_votes)) < 0.5:
results.append(None)
continue
results.append(best_person_name)
return results
def get_match_info(self, faces, max_distance=MAX_DISTANCE):
results = []
indices_s, distances_s = self.tree.query_radius(faces, r=max_distance, return_distance=True)
#print(indices_s, distances_s)
for indices, distances in zip(indices_s, distances_s):
inner_result = []
for index, distance in zip(indices, distances):
info = self.face_name_storage.get_info(index)
info['distance'] = distance
inner_result.append(info)
results.append(inner_result)
return results
def save_directory_to_model(directory, model_path):
people = load_people(directory)
serialized = []
index_map = []
files = []
test_data = []
for name, faces in people.items():
for face, filepath in faces:
#if random.random() < (TEST_PERCENT/100.0):
#test_data.append((name, face))
#else:
index_map.append(name)
serialized.append(face)
files.append(filepath)
people_list = get_people_list()
name_email = {p['name']: p['email'] for p in people_list}
fns = FaceNameStorage(index_map, serialized, files, name_email)
images_names = set(index_map)
known_names = set(name_email.keys())
if images_names != known_names:
print("In images but now known: %s" % str(images_names-known_names))
print("In known but not images: %s" % str(known_names-images_names))
raise Exception("Mismatch in known names!")
with open(model_path, "wb") as f:
print("wrote %s" % model_path)
pickle.dump(fns, f)
with open(model_path+".joblib", "wb") as f:
print("wrote %s" % model_path + ".joblib")
joblib.dump(fns, f)
return fns
def load_model(model_path):
with open(model_path, "rb") as f:
if model_path.endswith("joblib"):
return joblib.load(f)
else:
return pickle.load(f, encoding='latin-1')