-
Notifications
You must be signed in to change notification settings - Fork 0
/
CamVid.py
130 lines (93 loc) · 4.26 KB
/
CamVid.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
import keras_segmentation
import numpy as np
import cv2
import os
import glob
import tensorflow as tf
from tqdm import tqdm
from keras.preprocessing.image import img_to_array, load_img
from keras_segmentation import metrics
config = tf.ConfigProto()
config.gpu_options.allow_growth = True
sess = tf.Session(config = config)
# Check available GPU devices.
print("The following GPU devices are available: %s" % tf.test.gpu_device_name())
model = keras_segmentation.models.segnet.vgg_segnet(n_classes=32, input_height=704, input_width=928 )#716 956
model.train(
train_images = "dataset2/image/train",
train_annotations = "dataset2/label/train",
val_images=None,
val_annotations=None,
checkpoints_path = "./path_to_checkpoints/" ,
epochs = 1000,
batch_size = 1,
validate=False ,
val_batch_size=None ,
auto_resume_checkpoint=False ,
load_weights=None ,
steps_per_epoch=300,
optimizer_name='Adam'
)
all_prs = model.predict_multiple(
inp_dir = "dataset2/image/test",
out_dir = "path_to_predictions/",
checkpoints_path = "path_to_checkpoints/"
)
def evaluate_camvid():
# get seg (ground truth) and compare with pr (prediction)
# Get seg
segs_path = "dataset4/label/test"
segs = glob.glob( os.path.join(segs_path,"*.png") )
#print(len(segs))
imglabels = np.ndarray((len(segs),720, 960), dtype=np.uint8)
i=0
for x in range(len(segs)):
imgpath = segs[x]
pic_name = imgpath.split('/')[-1]
labelpath = "dataset2/label/" + pic_name.split('.')[0] + '.png'
label = cv2.imread(labelpath)
imglabels[i] = label[:,:,0]
if i % 100 == 0:
print('Creating testing images: {0}/{1} images'.format(i, len(segs)))
i += 1
#np.save('./segs_test.npy', imglabels)
i=0
orininal_w = 960
orininal_h = 720
n_classes = 32
ious=[]
precisions=[]
recalls=[]
for pr in tqdm(all_prs):
gt = imglabels[i]
pr = cv2.resize(pr, (orininal_w , orininal_h), interpolation=cv2.INTER_NEAREST)
#IoU
iou = metrics.get_iou( gt , pr , n_classes )
ious.append( iou )
# precision
precision = metrics.get_precision( gt , pr , n_classes )
precisions.append( precision )
# recall
recall = metrics.get_recall( gt , pr , n_classes )
recalls.append( recall )
i+=1
ious = np.array( ious )
precisions = np.array( precisions )
recalls = np.array( recalls )
print("Class wise IoU Class:{:d} / IoU:{:.2f}\n".format(11, np.mean(ious , axis=0 )[11]))
print("Class wise Precision Class:{:d} / Precision:{:.2f}\n".format(11, np.mean(precisions , axis=0 )[11]))
print("Class wise Recall Class:{:d} / Recall:{:.2f}\n".format(11, np.mean(recalls , axis=0 )[11]))
# print("Class wise IoU Class:{:d} / IoU:{:.2f}\n".format(12, np.mean(ious , axis=0 )[12]))
# print("Class wise Precision Class:{:d} / Precision:{:.2f}\n".format(12, np.mean(precisions , axis=0 )[12]))
# print("Class wise Recall Class:{:d} / Recall:{:.2f}\n".format(12, np.mean(recalls , axis=0 )[12]))
evaluate_camvid()
'''
# dataset1 (prepared dataset, 12 calsses)
python -m keras_segmentation train --checkpoints_path="path_to_checkpoints" --train_images="dataset1/images_prepped_train/" --train_annotations="dataset1/annotations_prepped_train/" --val_images="dataset1/images_prepped_test/" --val_annotations="dataset1/annotations_prepped_test/" --n_classes=12 --input_height=320 --input_width=640 --model_name="vgg_unet"
python -m keras_segmentation predict --checkpoints_path="./path_to_checkpoints/" --input_path="dataset2/image/test/" --output_path="path_to_predictions"
'''
'''
# dataset2 (CamVid, 32 classes)
python -m keras_segmentation train --checkpoints_path="path_to_checkpoints/" --train_images="dataset2/image/train/" --train_annotations="dataset2/trainId_label/train/" --val_images="dataset2/image/test/" --val_annotations="dataset2/trainId_label/test/" --n_classes=32 --input_height=736 --input_width=960 --model_name="vgg_unet" --epochs=20
python -m keras_segmentation predict --checkpoints_path="./path_to_checkpoints/" --input_path="dataset2/image/test/" --output_path="path_to_predictions"
'''