/
demo.py
152 lines (134 loc) · 6.26 KB
/
demo.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
from tensorflow.keras.preprocessing import image
from tensorflow.keras.models import model_from_json
import matplotlib.pyplot as plt
import numpy as np
import os, random, sys
from create_datagenerators import create_data_generators
import json
from sklearn.metrics import confusion_matrix
from data_preprocessing import resize_white, resize_black
from tensorflow.keras.utils import plot_model
from collections import Counter
SAVE_RESULRS_DIR = 'saved_models/'
RESULTS_FOLDER = SAVE_RESULRS_DIR + '/20190701_1148'
#20190612_1048
if (os.getcwd() == '/home/kalkami/translearn'or os.getcwd() == '/home/kalkami/translearn_cpu'):
#lhcpgpu1
TRAIN_DIR = '/data/IntelliGate/kalkami/DATASETS/carsStanford_all/train'
TEST_DIR = '/data/IntelliGate/kalkami/DATASETS/carsStanford_all/test'
TRAIN_DIR_TST = TRAIN_DIR
TEST_DIR_TST = TEST_DIR
elif (os.getcwd() == '/home/kamila/Desktop/InteliGate/CLASSIFICATION/VMMR/cars-classification-deep-learning'):
#local
TRAIN_DIR = '/media/kamila/System/Users/Kama/Documents/DATASETS/carsStanford_s/train'
TEST_DIR = '/media/kamila/System/Users/Kama/Documents/DATASETS/carsStanford_s/test'
TRAIN_DIR_TST = '/media/kamila/System/Users/Kama/Documents/DATASETS/carsStanford_all/train'
TEST_DIR_TST = '/media/kamila/System/Users/Kama/Documents/DATASETS/carsStanford_all/test'
TEST_DIR_TST = '/media/kamila/System/Users/Kama/Documents/DATASETS/CARS_GOOGLE_IMG/downloads'
else:
# default
TRAIN_DIR = 'DATASETS/Stanford_Dataset_sorted/train'
TEST_DIR ='DATASETS/Stanford_Dataset_sorted/test'
TRAIN_DIR_TST = TRAIN_DIR
TEST_DIR_TST = TEST_DIR
def load_image(img_path, input_shape, resize = False):
if resize:
img, pth = resize_black(input_shape[1], img_path, print_oldsize=False)
else:
img = image.load_img(img_path, target_size=input_shape)
img_tensor = image.img_to_array(img) # (height, width, channels)
img_tensor = np.expand_dims(img_tensor, axis=0) # (1, height, width, channels), add a dimension because the model expects this shape: (batch_size, height, width, channels)
img_tensor /= 255. # imshow expects values in the range [0, 1]
return img_tensor
def decode_predictions(preds, class_names, top=5):
results = []
for pred in preds:
top_indices = pred.argsort()[-top:][::-1]
result = [(class_names[i], pred[i]) for i in top_indices]
result.sort(key=lambda x: x[1], reverse=True)
results.append(result)
return results
def predict(img_path, model, input_shape, class_names, correct_class):
img_array = load_image(img_path, input_shape)
preds = model.predict(img_array)
predictions = decode_predictions(preds, class_names)
top1_pred = predictions[0][0]
print(predictions)
img_org = image.load_img(img_path)
fig, axs = plt.subplots(1,2)
axs[0].set_title(correct_class)
axs[0].imshow(img_org)
#axs[0].axis('off')
axs[1].set_title(str(top1_pred))
axs[1].imshow(img_array[0])
plt.show()
# Decode the output of the VGG16 model.
#pred_decoded = decode_predictions(pred)[0]
## Print the predictions.
#for code, name, score in pred_decoded:
#print("{0:>6.2%} : {1}".format(score, name))
def load_model(results_folder, show_accuracy=False):
# load json and create model
json_file = open(results_folder + '/model.json', 'r')
loaded_model_json = json_file.read()
json_file.close()
loaded_model = model_from_json(loaded_model_json)
# load weights into new model
loaded_model.load_weights(results_folder + "/weights.best.hdf5")
input_shape = loaded_model.layers[0].output_shape[1:3]
print("Loaded model from disk")
return loaded_model, input_shape
def perform_pred(car_class, results_folder=RESULTS_FOLDER, test_dir=TEST_DIR_TST, img_pth=None):
HYPERPARAMS_FILE = results_folder+ '/hyperparams.json'
with open(HYPERPARAMS_FILE, "r") as read_file:
data = json.load(read_file)
HYPERPARAMS = data['hyperparameters'][0]
BATCHSIZE = HYPERPARAMS['BATCHSIZE']
if img_pth is None:
# randomly select an image from defined class
test_dir_full = test_dir + '/' + car_class
test_img = test_dir_full + '/' + random.choice(os.listdir(test_dir_full))
else:
test_img = img_pth
loaded_model, input_shape=load_model(results_folder)
print (input_shape[1])
if os.path.exists(results_folder+'/class_names.txt'):
class_names = []
# open file and read the content in a list
with open(results_folder+'/class_names.txt', 'r') as filehandle:
for line in filehandle:
current_line = line[:-1]
class_names.append(current_line)
else:
generator_train, generator_test = create_data_generators(input_shape, BATCHSIZE,
TRAIN_DIR, TEST_DIR,
save_augumented=None,
plot_imgs = False)
class_names = list(generator_train.class_indices.keys())
print(class_names)
with open(results_folder+'/class_names.txt', 'w') as filehandle:
for listitem in class_names:
filehandle.write('%s\n' % listitem)
predict(test_img, loaded_model, input_shape, class_names, car_class)
if __name__ == "__main__":
#img_test = '../test_imgs/f.png'
# Audi A5 Coupe 2012
print(sys.argv)
if len(sys.argv) == 1:
print('Too few arguments.')
elif len(sys.argv) == 2:
perform_pred(sys.argv[1])
elif len(sys.argv) == 3:
if str(sys.argv[2]).endswith('.jpg') or str(sys.argv[2]).endswith('.png'):
perform_pred(sys.argv[1], img_pth=sys.argv[2])
else:
perform_pred(sys.argv[1], results_folder=sys.argv[2])
elif len(sys.argv) == 4:
if str(sys.argv[3]).endswith('.jpg') or str(sys.argv[3]).endswith('.png'):
perform_pred(sys.argv[1], results_folder=sys.argv[2],
img_pth=sys.argv[3])
else:
perform_pred(sys.argv[1], results_folder=sys.argv[2],
test_dir=sys.argv[3])
else:
print('Too much arguments.')