-
Notifications
You must be signed in to change notification settings - Fork 1
/
data.py
137 lines (119 loc) · 5.15 KB
/
data.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
import os
from PIL import Image
import torch.utils.data as data
import torchvision.transforms as transforms
import random
import numpy as np
class SalObjDataset(data.Dataset):
# def __init__(self, image_root, gt_root, edge_root, trainsize):
def __init__(self, image_root, gt_root, trainsize):
self.trainsize = trainsize
self.images = [image_root + f for f in os.listdir(image_root) if f.endswith('.jpg')]
self.gts = [gt_root + f for f in os.listdir(gt_root) if f.endswith('.jpg')
or f.endswith('.png')]
# self.edges = [gt_root + f for f in os.listdir(edge_root) if f.endswith('.jpg')
# or f.endswith('.png')]
self.images = sorted(self.images)
# self.depths = sorted(self.depths)
self.gts = sorted(self.gts)
# self.edges = sorted(self.edges)
self.filter_files()
self.size = len(self.images)
self.img_transform = transforms.Compose([
transforms.Resize((self.trainsize, self.trainsize)),
transforms.ToTensor(),
transforms.Normalize([0.485, 0.456, 0.406], [0.229, 0.224, 0.225])])
self.gt_transform = transforms.Compose([
transforms.Resize((self.trainsize, self.trainsize)),
transforms.ToTensor()])
def __getitem__(self, index):
image = self.rgb_loader(self.images[index])
gt = self.binary_loader(self.gts[index])
image = self.img_transform(image)
gt = self.gt_transform(gt)
return image, gt
def filter_files(self):
assert len(self.images) == len(self.gts)
images = []
# depths = []
gts = []
# edges = []
# for img_path, gt_path, edge_path in zip(self.images, self.gts, self.edges):
for img_path, gt_path in zip(self.images, self.gts):
img = Image.open(img_path)
gt = Image.open(gt_path)
# edge = Image.open(edge_path)
if img.size == gt.size:
images.append(img_path)
gts.append(gt_path)
# edges.append(edge_path)
self.images = images
self.gts = gts
# self.edges = edges
def rgb_loader(self, path):
with open(path, 'rb') as f:
img = Image.open(f)
return img.convert('RGB')
def binary_loader(self, path):
with open(path, 'rb') as f:
img = Image.open(f)
# return img.convert('1')
return img.convert('L')
# def resize(self, img, gt, edge):
def resize(self, img, gt):
assert img.size == gt.size
w, h = img.size
if h < self.trainsize or w < self.trainsize:
h = max(h, self.trainsize)
w = max(w, self.trainsize)
# return img.resize((w, h), Image.BILINEAR), gt.resize((w, h), Image.NEAREST), edge.resize((w, h), Image.NEAREST)
return img.resize((w, h), Image.BILINEAR), gt.resize((w, h), Image.NEAREST)
else:
# return img, gt, edge
return img, gt
def __len__(self):
return self.size
# def get_loader(image_root, gt_root, edge_root, batchsize, trainsize, shuffle=True, num_workers=12, pin_memory=True):
def get_loader(image_root, gt_root, batchsize, trainsize, shuffle=True, num_workers=0, pin_memory=True):
# dataset = SalObjDataset(image_root, gt_root, edge_root, trainsize)
dataset = SalObjDataset(image_root, gt_root, trainsize)
data_loader = data.DataLoader(dataset=dataset,
batch_size=batchsize,
shuffle=shuffle,
num_workers=num_workers,
pin_memory=pin_memory)
return data_loader
class test_dataset:
def __init__(self, image_root, gt_root, testsize):
self.testsize = testsize
self.images = [image_root + f for f in os.listdir(image_root) if f.endswith('.jpg')]
self.gts = [gt_root + f for f in os.listdir(gt_root) if f.endswith('.jpg')
or f.endswith('.png')]
self.images = sorted(self.images)
# self.depths = sorted(self.depths)
self.gts = sorted(self.gts)
self.img_transform = transforms.Compose([
transforms.Resize((self.testsize, self.testsize)),
transforms.ToTensor(),
transforms.Normalize([0.485, 0.456, 0.406], [0.229, 0.224, 0.225])])
self.gt_transform = transforms.ToTensor()
self.size = len(self.images)
self.index = 0
def load_data(self):
image = self.rgb_loader(self.images[self.index])
image = self.img_transform(image).unsqueeze(0)
gt = self.binary_loader(self.gts[self.index])
name = self.images[self.index].split('/')[-1]
if name.endswith('.jpg'):
name = name.split('.jpg')[0] + '.png'
self.index += 1
# self.index = self.index % self.size
return image, gt, name
def rgb_loader(self, path):
with open(path, 'rb') as f:
img = Image.open(f)
return img.convert('RGB')
def binary_loader(self, path):
with open(path, 'rb') as f:
img = Image.open(f)
return img.convert('L')