-
Notifications
You must be signed in to change notification settings - Fork 0
/
face_recognition.py
81 lines (73 loc) · 3.84 KB
/
face_recognition.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
from data_process import pickle_read
from data_generator import DataGenerator, load_all
from classifier_models import get_classifier
from sklearn.metrics import precision_recall_fscore_support
import numpy as np
from tensorflow.keras.models import Sequential, Model, load_model
from tensorflow.keras.layers import Dense,Dropout,Softmax,Flatten,Activation,BatchNormalization
import tensorflow as tf
class FaceRecongizer():
def __init__(self):
self.labels = {-1: 'unknown', 0: 'ashwin', 1: 'tessa'} # TODO: change the labels
self.load_pretrained = True
self.path_save_model = "./data/models/face_recognizer/face_recognizer.h5"
self.get_model()
def get_model(self):
if self.load_pretrained:
self.model = load_model(self.path_save_model)
return
self.model = Sequential()
self.model.add(Dense(units=100,input_dim=128,kernel_initializer='glorot_uniform'))
self.model.add(BatchNormalization())
self.model.add(Activation('relu'))
self.model.add(Dropout(0.3))
self.model.add(Dense(units=10,kernel_initializer='glorot_uniform'))
self.model.add(BatchNormalization())
self.model.add(Activation('relu'))
self.model.add(Dropout(0.2))
self.model.add(Dense(units=len(self.labels),kernel_initializer='he_uniform'))
self.model.add(Activation('softmax'))
self.model.compile(loss=tf.keras.losses.SparseCategoricalCrossentropy(),optimizer='nadam',metrics=['accuracy'])
def train(self):
# Load meta data of dataset
train_set = pickle_read("./data/image_recognition/processed/train.pkl")
valid_set = pickle_read("./data/image_recognition/processed/valid.pkl")
test_set = pickle_read("./data/image_recognition/processed/test.pkl")
self.get_model()
# Train with data generator
gen_train = DataGenerator(dataset=train_set, batch_size=8, shuffle=True)
gen_valid = DataGenerator(dataset=valid_set, batch_size=8, shuffle=True)
gen_test = DataGenerator(dataset=test_set, batch_size=8)
history = self.model.fit_generator(epochs=3, generator=gen_train, validation_data=gen_valid)
scores = self.model.evaluate_generator(gen_test, 4)
print("scores", scores[1])
# # Train without data generator
# x_train, y_train = load_all(train_set)
# x_valid, y_valid = load_all(valid_set)
# x_test, y_test = load_all(test_set)
# history = self.model.fit(x=x_train, y=y_train, batch_size=128,
# epochs=3, validation_data=(x_valid, y_valid), shuffle=True)
# scores = self.model.evaluate(x=x_test, y=y_test, batch_size=128)
self.model.save(self.path_save_model)
def evaluate(self):
test_set = pickle_read("./data/image_recognition/processed/test.pkl")
# gen_test = DataGenerator(dataset=test_set, batch_size=4)
# data generator alternate
x_test, y_test = load_all(test_set)
preds = self.model.predict(x_test)
preds = np.argmax(preds, axis=-1)
return precision_recall_fscore_support(y_true=y_test, y_pred=preds, average="micro")
def predict(self, embedding, threshold, verbose):
embedding = np.expand_dims(embedding, axis=0)
pred = self.model.predict(embedding)
if verbose:
print("Recognition probability:", np.max(pred))
print("Person:", self.labels[np.argmax(pred, axis=-1)[0]])
print("Is valid:", np.max(pred) > threshold)
if np.max(pred) > threshold:
return self.labels[np.argmax(pred, axis=-1)[0]], np.max(pred) > threshold
return self.labels[-1], np.max(pred) > threshold
if __name__ == "__main__":
face_recognizer = FaceRecongizer()
face_recognizer.train()
print("precision, recall, fscore, support:", face_recognizer.evaluate())