SPNet/extract_features.py at main · abdkhanstd/SPNet · GitHub

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

"""
This script generates extracted features for each patient, which other
models make use of.
"""

"""
Train all the basline networks
"""
import os
#os.environ['CUDA_VISIBLE_DEVICES'] ='5'
###### Collect all garbage #######
import sys
stderr = sys.stderr
sys.stderr = open(os.devnull, 'w')
import keras
sys.stderr = stderr
###################################

import tensorflow as tf
from keras.backend.tensorflow_backend import set_session
gpu_options = tf.GPUOptions(per_process_gpu_memory_fraction=0.2)
sess = tf.Session(config=tf.ConfigProto(
  allow_soft_placement=True, log_device_placement=True))


import numpy as np
import os.path
from utils.data import DataSet
from tqdm import tqdm

# Set defaults.
seq_length = 16                  #                     H
extractor_arch='IV3' #Options (resnet50, IV3,vgg16,vgg19,IR2,resnet152,densenet169,xception,efficientnetb7)
target='views'


class_limit = None
ndir=os.path.join('data','sequences_'+extractor_arch)
if not os.path.exists(ndir):
    os.makedirs(ndir)
# Get the dataset.
data = DataSet(seq_length=seq_length, class_limit=class_limit,target=target)
# get the model.
if extractor_arch=='IV3':
    from  extractors.extractor_iv3 import Extractor
    model = Extractor()
if extractor_arch=='resnet50':
    from  extractors.extractor_resnet50 import Extractor
    model = Extractor()
if extractor_arch=='vgg16':
    from  extractors.extractor_vgg16 import Extractor
    model = Extractor()
if extractor_arch=='IR2':
    from  extractors.extractor_IR2 import Extractor
    model = Extractor()
if extractor_arch=='resnet152':
    from  extractors.extractor_resnet152 import Extractor
    model = Extractor()
if extractor_arch=='densenet169':
    from  extractors.extractor_densenet169 import Extractor
    model = Extractor()
if extractor_arch=='vgg19':
    from  extractors.extractor_vgg19 import Extractor
    model = Extractor()
if extractor_arch=='xception':
    from  extractors.extractor_xception import Extractor
    model = Extractor()
if extractor_arch=='efficientnetb7':
    from  extractors.extractor_efficientnetb7 import Extractor
    model = Extractor()


# Loop through data.
pbar = tqdm(total=len(data.data))
for video in data.data:

    #trim the file name
    tmp=video[2].split('/')
    tmp=tmp[1].split('.')
    fname=tmp[0]


    # Get the path to the sequence for this video.
    path = os.path.join('data', 'sequences_'+extractor_arch, fname + '-' + str(seq_length) + \
        '-features_'+target)  # numpy will auto-append .npy

    # Check if we already have it.
    if os.path.isfile(path + '.npy'):
        pbar.update(1)
        continue

    # Get the frames for this video.
    frames = data.get_frames_for_sample(video)

    # Now downsample to just the ones we need.
    frames = data.rescale_list(frames, seq_length)

    # Now loop through and extract features to build the sequence.
    sequence = []
    for image in frames:
        features = model.extract(image)
        features=np.squeeze(features)         #remove this line for python 2.7
        sequence.append(features)

    # Save the sequence.
    np.save(path, sequence)

    pbar.update(1)

pbar.close()