-
Notifications
You must be signed in to change notification settings - Fork 0
/
test_tflite.py
200 lines (163 loc) · 7.34 KB
/
test_tflite.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
import numpy as np
import tensorflow as tf
import cv2 as cv
import glob
import re
MODELS = [
"MnasNet_E25_B8_A0.5_DEPTH5_Adam0.0001_AUGFULL_SHUFFLE_float16.tflite",
]
MODEL_NAME = "EfficientNetB0_E25_B4_AUGFULL_SHUFFLE_float16.tflite"
MODEL_PATH = f".\\nets_tflite\\{MODEL_NAME}"
DICT = {0: "Neutral", 1: "Happiness", 2: "Sadness", 3: "Surprise", 4: "Fear", 5: "Disgust", 6: "Anger", 7: "Contempt", 8: "None", 9: "Uncertain", 10: "No-Face"}
TEST_IMAGES_PATH = "C:\\Users\\Vojta\\DiplomaProjects\\AffectNet\\val_set\\images\\"
TEST_LABELS_PATH = "C:\\Users\\Vojta\\DiplomaProjects\\AffectNet\\val_set\\all_labels_exp.npy"
TEST_ARO_LABELS_PATH = "C:\\Users\\Vojta\\DiplomaProjects\\AffectNet\\val_set\\all_labels_aro.npy"
TEST_VAL_LABELS_PATH = "C:\\Users\\Vojta\\DiplomaProjects\\AffectNet\\val_set\\all_labels_val.npy"
#TEST_IMAGES_PATH = "/sp1/val_set/images/"
#TEST_LABELS_PATH = "/sp1/val_set/all_labels_exp.npy"
TEST_ONE_IMG = False
WEBCAM = False
REGRESSION = False
def testOneImage():
# Load TFLite model and allocate tensors.
interpreter = tf.lite.Interpreter(model_path = MODEL_PATH)
interpreter.allocate_tensors()
# Get input and output tensors.
input_details = interpreter.get_input_details()
output_details = interpreter.get_output_details()
# Test model on image
#img = cv.imread("test_imgs\\angry.jpg", 1)
#img = cv.imread("test_imgs\\happy.jpg", 1)
#img = cv.imread("test_imgs\\happy2.jpg", 1)
#img = cv.imread("test_imgs\\sad.jpg", 1)
img = cv.imread("test_imgs\\surprise.jpg", 1)
cascade = cv.CascadeClassifier("haarcascade_frontalface_default.xml")
rect = cascade.detectMultiScale(img, 1.3, 3)[0]
img = img[rect[1]:(rect[1] + rect[3]), rect[0]:(rect[0] + rect[2])]
img = cv.resize(img, (224, 224))
img = img.reshape(1, 224, 224, 3)
img = np.array(img, dtype = np.float32)
#img = img / 255.0
interpreter.set_tensor(input_details[0]['index'], img)
interpreter.invoke()
output = interpreter.get_tensor(output_details[0]["index"])[0]
#print(output)
print(DICT[np.argmax(output)], output[np.argmax(output)], output)
def webcamTest():
cascade = cv.CascadeClassifier("haarcascade_frontalface_default.xml")
capture = cv.VideoCapture(0)
while (True):
ret, image = capture.read()
rectangles = cascade.detectMultiScale(image, 1.3, 3)
# Load TFLite model and allocate tensors.
interpreter = tf.lite.Interpreter(model_path = MODEL_PATH)
interpreter.allocate_tensors()
# Get input and output tensors.
input_details = interpreter.get_input_details()
output_details = interpreter.get_output_details()
for rect in rectangles:
img = image[rect[1]:(rect[1] + rect[3]), rect[0]:(rect[0] + rect[2])]
img = cv.resize(img, (224, 224))
img = img.reshape(1, 224, 224, 3)
img = np.array(img, dtype = np.float32)
interpreter.set_tensor(input_details[0]['index'], img)
interpreter.invoke()
output = interpreter.get_tensor(output_details[0]["index"])[0]
emotion = DICT[np.argmax(output)]
prediction = output[np.argmax(output)]
print(f"{(prediction * 100):.3f} %\t{emotion} ", end = "\r")
cv.rectangle(image, rect, (0, 0, 0), 3)
cv.rectangle(image, rect, (255, 255, 255), 1)
cv.putText(image, emotion, (rect[1], rect[0]), cv.FONT_HERSHEY_SIMPLEX, 1, (0, 0, 0), 3, 2)
cv.putText(image, emotion, (rect[1], rect[0]), cv.FONT_HERSHEY_SIMPLEX, 1, (255, 255, 255), 1, 2)
cv.imshow("webcam", image)
if cv.waitKey(1) & 0xFF == ord('q'):
break
def testValDataset():
labels = np.load(TEST_LABELS_PATH)
predictions = []
images_paths_list = glob.glob(TEST_IMAGES_PATH + "*.jpg")
images_paths_list.sort(key = natural_keys)
errors = 0
# Load TFLite model and allocate tensors.
interpreter = tf.lite.Interpreter(model_path = MODEL_PATH, num_threads = 8)
interpreter.allocate_tensors()
# Get input and output tensors.
input_details = interpreter.get_input_details()
output_details = interpreter.get_output_details()
print(MODEL_NAME)
for i in range(len(images_paths_list)):
img_path = images_paths_list[i]
img = cv.imread(img_path, 1)
img = img.reshape(1, 224, 224, 3)
img = np.array(img, dtype = np.float32)
interpreter.set_tensor(input_details[0]['index'], img)
interpreter.invoke()
prediction = interpreter.get_tensor(output_details[0]["index"])[0]
predictions.append(np.argmax(prediction))
if np.argmax(prediction) != labels[i]:
errors += 1
evaluation = (1 - (errors / (i + 1))) * 100
print(f"{i} / {len(images_paths_list)}\t\tSuccess rate: {evaluation:.3f} % ", end = "\r")
evaluation = (1 - (errors / (len(images_paths_list)))) * 100
print(f"\nImages: {len(images_paths_list)}\nErrors: {errors}\nSuccess rate: {evaluation:.3f} %")
print(f"\nConfusion matrix:\n {tf.math.confusion_matrix(labels, predictions)}")
f = open(".\\nets_tflite\\stats_classifier.txt", "a")
#f = open("./nets_tflite/stats_classifier.txt", "a")
f.write(f"{MODEL_NAME}\nSuccess rate: {evaluation:.3f} %\nConfusion matrix:\n{tf.math.confusion_matrix(labels, predictions)}\n\n")
f.close()
def testValDatasetRegression():
labels_aro = np.load(TEST_ARO_LABELS_PATH)
labels_val = np.load(TEST_VAL_LABELS_PATH)
images_paths_list = glob.glob(TEST_IMAGES_PATH + "*.jpg")
images_paths_list.sort(key = natural_keys)
labels = [[labels_aro[i], labels_val[i]] for i in range(len(images_paths_list))]
RMSE_avg_aro = 0
RMSE_avg_val = 0
# Load TFLite model and allocate tensors.
interpreter = tf.lite.Interpreter(model_path = MODEL_PATH, num_threads = 8)
interpreter.allocate_tensors()
# Get input and output tensors.
input_details = interpreter.get_input_details()
output_details = interpreter.get_output_details()
print(MODEL_NAME)
for i in range(len(images_paths_list)):
img_path = images_paths_list[i]
img = cv.imread(img_path, 1)
img = img.reshape(1, 224, 224, 3)
img = np.array(img, dtype = np.float32)
interpreter.set_tensor(input_details[0]['index'], img)
interpreter.invoke()
aro_pred, val_pred = interpreter.get_tensor(output_details[0]["index"])[0]
aro_label, val_label = labels[i]
RMSE_avg_aro += (aro_pred - aro_label) ** 2
RMSE_avg_val += (val_pred - val_label) ** 2
print(f"{i} / {len(images_paths_list)}\t\tArousal avg RMSE: {(np.sqrt((1 / (i + 1)) * RMSE_avg_aro)):.4f}\t\tValence avg RMSE: {(np.sqrt((1 / (i + 1)) * RMSE_avg_val)):.4f} ", end = "\r")
RMSE_avg_aro = np.sqrt((1 / len(images_paths_list)) * RMSE_avg_aro)
RMSE_avg_val = np.sqrt((1 / len(images_paths_list)) * RMSE_avg_val)
print("\n")
print(f"{MODEL_NAME}\nImages: {len(images_paths_list)}\nArousal average RMSE: {RMSE_avg_aro:.4f}\nValence average RMSE: {RMSE_avg_val:.4f}\nAverage total RMSE: {((RMSE_avg_aro + RMSE_avg_val) / 2):.4f}")
f = open(".\\nets_tflite\\stats_regressor.txt", "a")
f.write(f"{MODEL_NAME}\nImages: {len(images_paths_list)}\nArousal average RMSE: {RMSE_avg_aro:.8f}\nValence average RMSE: {RMSE_avg_val:.8f}\nAverage total RMSE: {((RMSE_avg_aro + RMSE_avg_val) / 2):.8f}")
f.write("\n\n")
f.close()
def atoi(text):
return int(text) if text.isdigit() else text
def natural_keys(text):
return [ atoi(c) for c in re.split(r'(\d+)', text) ]
if __name__ == '__main__':
if TEST_ONE_IMG:
testOneImage()
elif WEBCAM:
webcamTest()
elif REGRESSION:
for i in MODELS:
MODEL_NAME = i
MODEL_PATH = f".\\nets_tflite\\{MODEL_NAME}"
testValDatasetRegression()
else:
for i in MODELS:
MODEL_NAME = i
MODEL_PATH = f".\\nets_tflite\\{MODEL_NAME}"
#MODEL_PATH = f"./nets_tflite/{MODEL_NAME}"
testValDataset()