-
Notifications
You must be signed in to change notification settings - Fork 2
/
facenet_utils.py
145 lines (120 loc) · 5.98 KB
/
facenet_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
# Compute the 128D vector that describes the face in img identified by
# shape. In general, if two face descriptor vectors have a Euclidean
# distance between them less than 0.6 then they are from the same
# person, otherwise they are from different people.
from __future__ import division
from __future__ import print_function
import os
import sys
import math
import cv2
import tqdm
import pickle
import argparse
import numpy as np
import tensorflow as tf
from sklearn import metrics
from scipy import interpolate
from scipy.optimize import brentq
import facenet
from MTCNNWrapper import MTCNNWrapper
class FaceUtil:
def __init__(self, data_path='images', model_dir='model/20180402-114759'):
self.data_path = data_path
self.model_dir = model_dir
self.mtcnn = MTCNNWrapper()
self.images_placeholder = None
self.embeddings = None
self.phase_train_placeholder=None
self.embedding_size=None
self.image_size = 160
self.margin=44
self.detect_multiple_faces = False
def convert_to_embedding(self, single=False, img_path=None):
extracted = []
with tf.Graph().as_default():
with tf.Session() as sess:
self.sess = sess
# Load the model
facenet.load_model(self.model_dir)
# Get input and output tensors
images_placeholder = tf.get_default_graph().get_tensor_by_name("input:0")
self.images_placeholder = tf.image.resize_images(images_placeholder,(self.image_size, self.image_size))
self.embeddings = tf.get_default_graph().get_tensor_by_name("embeddings:0")
self.phase_train_placeholder = tf.get_default_graph().get_tensor_by_name("phase_train:0")
self.embedding_size = self.embeddings.get_shape()[1]
if not single:
for filename in os.listdir(self.data_path):
#print ("filename -> ", filename)
file_path = os.path.join(self.data_path, filename)
img = cv2.imread(file_path, 1)
bounding_boxes, points = self.mtcnn.get_result_dict(image=img)
faces = self.get_faces(img, bounding_boxes, points, filename)
extracted.append(faces)
return extracted
else:
img = cv2.imread(img_path, 1)
bounding_boxes, points = self.mtcnn.get_result_dict(image=img)
faces = self.get_faces(img, bounding_boxes, points, img_path)
return faces
def get_faces(self, img, bounding_boxes, points, filename):
faces = []
nrof_faces = bounding_boxes.shape[0]
#print ('bounding_boxes.shape', bounding_boxes.shape[0])
#print("No. of faces detected: {}".format(nrof_faces))
if nrof_faces>0:
det = bounding_boxes[:,0:4]
det_arr = []
img_size = np.asarray(img.shape)[0:2]
if nrof_faces>1:
if self.detect_multiple_faces:
for i in range(nrof_faces):
det_arr.append(np.squeeze(det[i]))
else:
bounding_box_size = (det[:,2]-det[:,0])*(det[:,3]-det[:,1])
img_center = img_size / 2
offsets = np.vstack([ (det[:,0]+det[:,2])/2-img_center[1], (det[:,1]+det[:,3])/2-img_center[0] ])
offset_dist_squared = np.sum(np.power(offsets,2.0),0)
index = np.argmax(bounding_box_size-offset_dist_squared*2.0) # some extra weight on the centering
det_arr.append(det[index,:])
else:
det_arr.append(np.squeeze(det))
for i, det in enumerate(det_arr):
det = np.squeeze(det)
bb = np.zeros(4, dtype=np.int32)
bb[0] = np.maximum(det[0]-self.margin/2, 0)
bb[1] = np.maximum(det[1]-self.margin/2, 0)
bb[2] = np.minimum(det[2]+self.margin/2, img_size[1])
bb[3] = np.minimum(det[3]+self.margin/2, img_size[0])
cropped = img[bb[1]:bb[3],bb[0]:bb[2],:]
resized = cv2.resize(cropped, (self.image_size, self.image_size),interpolation=cv2.INTER_CUBIC)
prewhitened = facenet.prewhiten(resized)
faces.append({'name': filename,'rect':[bb[0],bb[1],bb[2],bb[3]],'embedding':self.get_embedding(prewhitened)})
return faces
def get_embedding(self, processed_img):
reshaped = processed_img.reshape(-1, self.image_size, self.image_size, 3)
feed_dict = {self.images_placeholder:reshaped, self.phase_train_placeholder:False }
feature_vector = self.sess.run(self.embeddings, feed_dict=feed_dict)
return feature_vector
def get_eucledian_dist_list(self, emb_list, embedding):
dist_list = []
for emb in emb_list:
dist = np.sqrt(np.sum(np.square(np.subtract(emb[0]['embedding'], embedding[0]['embedding']))))
#print('distance from {}'.format(emb[0]['name'].split('.')[0]))
#print(' %1.4f ' % dist, end='')
#print("\n")
dist_list.append([emb[0]['name'], dist])
return dist_list
def get_eucledian_dist(self, enc1, enc2):
dist = np.sqrt(np.sum(np.square(np.subtract(enc1[0]['embedding'], enc2[0]['embedding']))))
return dist
if __name__ == '__main__':
face_util = FaceUtil()
# if pickle file does not exists, generate it
if not os.path.exists('dbfile'):
encodings = face_util.convert_to_embedding()
# read the pickle file
encodings_list = face_util.load_data()
embedding = face_util.convert_to_embedding(single=True, img_path='sarah.JPG')
encodings_list = face_util.load_data()
face_util.get_eucledian_dist_list(encodings_list, embedding)