TransOrga/dataset.py at main · LittleQBerry/TransOrga · 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
114
115
116
117
118
119
120
import os
import torch
from PIL import Image
from torchvision import transforms
from torch import nn
import numpy as np
import cv2
class Dataset(torch.utils.data.Dataset):
    '''
    train, test, validatiaon
    '''
    def __init__(self, path, transform=None, mode='train', val_path = None):
        self.img_path = path
        self.mask_path = path
        self.transform = transform
        self.mode = mode
        if self.mode == 'train':
            self.img_path =self.img_path +"/training/images/"
            self.mask_path = self.mask_path +"/training/segmentations/"
            self.edge_path ='/edge_result/training/'
            self.sr_path = '/SR_results/training/'
            train_file = os.listdir(self.img_path)
            files=train_file
        elif self.mode == 'validation':
            self.img_path =self.img_path +"/validation/images/"
            self.mask_path = self.mask_path +"/validation/segmentations/"
            self.edge_path = '/edge_result/validation/'
            self.sr_path ='/SR_results/validation/'
            valid_file = os.listdir(self.img_path)
            files =valid_file
        elif self.mode == 'test':
            self.img_path =self.img_path +"/testing/images/"
            self.mask_path = self.mask_path +"/testing/segmentations/"
            self.edge_path = '/edge_result/testing/'
            self.sr_path = '/SR_results/testing/'
            test_file = os.listdir(self.img_path)
            files=test_file
        self.files = files
        #self.log = log


    def __getitem__(self,index):
        if self.mode == 'train' :
            name,_ = os.path.splitext(self.files[index])
            img = Image.open(self.img_path+self.files[index].strip()).convert('L')
            mask = Image.open(self.mask_path+self.files[index].strip()).convert('L')
            edge = Image.open(self.edge_path+self.files[index].strip()).convert('L')
            seed = torch.random.seed()
            sr =torch.load("{}/{}.pt".format(self.sr_path,name))
            sr =sr[0]
            torch.random.manual_seed(seed)
            img = self.transform(img)
            sr =self.transform(sr)
            #sr = transforms.ToTensor()(sr)
            img = transforms.ColorJitter(brightness=0.5,contrast=0.5,hue=0.5)(img)
            img = transforms.ToTensor()(img)
            torch.random.manual_seed(seed)
            mask = self.transform(mask)
            mask =transforms.ToTensor()(mask)
            mask = torch.where(mask>0.0,1.0,0.0)

            edge = self.transform(edge)
            edge = transforms.ToTensor()(edge)


            file_name = self.files[index].strip()
            img_r =torch.cat([img,sr],dim=0)
            return img_r, mask, file_name
        elif self.mode == 'validation':

            name,_ = os.path.splitext(self.files[index])
            img = Image.open(self.img_path+self.files[index].strip()).convert('L')
            mask = Image.open(self.mask_path+self.files[index].strip()).convert('L')
            edge = Image.open(self.edge_path+self.files[index].strip()).convert('L')
            seed = torch.random.seed()
            sr =torch.load("{}/{}.pt".format(self.sr_path,name))
            sr =sr[0]

            torch.random.manual_seed(seed)
            img = self.transform(img)
            sr = self.transform(sr)
            img = transforms.ToTensor()(img)
            torch.random.manual_seed(seed)
            mask = self.transform(mask)
            mask =transforms.ToTensor()(mask)
            mask = torch.where(mask>0.0,1.0,0.0)

            edge = self.transform(edge)
            edge =transforms.ToTensor()(edge)
            file_name = self.files[index].strip()
            img_r =torch.cat([img,sr],dim=0)
            return img_r, mask, file_name
        elif self.mode == 'test':
            name,_ = os.path.splitext(self.files[index])
            img =cv2.imread(self.img_path+self.files[index].strip(),0)
            img =Image.fromarray(img.astype('uint8'))
            mask = Image.open(self.mask_path+self.files[index].strip()).convert('L')
            edge = Image.open(self.edge_path+self.files[index].strip()).convert('L')
            seed = torch.random.seed()
            sr =torch.load("{}/{}.pt".format(self.sr_path,name))
            sr =sr[0]

            torch.random.manual_seed(seed)
            img = self.transform(img)
            sr = self.transform(sr)
            img = transforms.ToTensor()(img)
            torch.random.manual_seed(seed)
            mask = self.transform(mask)
            mask =transforms.ToTensor()(mask)
            mask = torch.where(mask>0.0,1.0,0.0)
            print(img.shape)
            edge = self.transform(edge)
            edge =transforms.ToTensor()(edge)
            file_name = self.files[index].strip()
            img_r =torch.cat([img,sr],dim=0)
            return img_r, mask, file_name

    def __len__(self):
        return len(self.files)