-
Notifications
You must be signed in to change notification settings - Fork 1
/
split_data_cub.py
97 lines (70 loc) · 2.8 KB
/
split_data_cub.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
import os
import shutil
import sys
import random
import errno
def load_class_names(dataset_path):
classes = {}
with open(os.path.join(dataset_path, "classes.txt")) as f:
for line in f:
(k, c) = line.split()
classes[int(k)] = c
return classes
def load_image_labels(dataset_path):
labels = {}
with open(os.path.join(dataset_path, "image_class_labels.txt")) as f:
for line in f:
(k, c) = line.split()
labels[int(k)] = int(c)
return labels
def load_image_paths(dataset_path, path_prefix=''):
paths = {}
with open(os.path.join(dataset_path, 'images.txt')) as f:
for line in f:
(k, p) = line.split()
path = os.path.join(path_prefix, p)
paths[int(k)] = path
return paths
def split_each_class(class_names, image_labels, split_train=0.60, split_val=0.20, split_test=0.20):
splits = {}
for c in class_names.keys():
# Find all images with label c
class_images = [k for k,v in image_labels.items() if v == c]
# Count images with label c
class_count = len(class_images)
# Split 60/20/20 train/val/test
train_count = round(class_count * split_train)
val_count = round(class_count * split_val)
test_count = round(class_count * split_test)
image_indices = list(range(class_count))
random.shuffle(image_indices)
train_indices = image_indices[0:train_count]
val_indices = image_indices[train_count:train_count+val_count]
test_indices = image_indices[train_count+val_count:]
for i in train_indices:
splits[class_images[i]] = 0
for i in val_indices:
splits[class_images[i]] = 1
for i in test_indices:
splits[class_images[i]] = 2
return splits
def copy_by_split(class_splits, image_paths, source_base, destination_base):
folders = {0: "train", 1: "val", 2: "test"}
for k,v in class_splits.items():
old_path = os.path.join(source_base, image_paths[k])
new_path = os.path.join(destination_base, folders[v], image_paths[k])
try:
shutil.copy2(old_path, new_path)
except IOError as e:
if e.errno != errno.ENOENT:
raise
os.makedirs(os.path.dirname(new_path))
shutil.copy2(old_path, new_path)
dataset_path = "C:/Users/dlohr/Downloads/cv-bird-classification/CUB_200_2011"
image_path_prefix = "images"
destination_path = "./data/cub-200-2011"
class_names = load_class_names(dataset_path)
image_labels = load_image_labels(dataset_path)
image_paths = load_image_paths(dataset_path, image_path_prefix)
class_splits = split_each_class(class_names, image_labels, 0.60, 0.20, 0.20)
copy_by_split(class_splits, image_paths, dataset_path, destination_path)