TPFNet/dataset.py at main · CandleLabAI/TPFNet · 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
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
import cv2
import numpy as np
import torch
from torch.utils.data import Dataset, DataLoader
from cfg import Config as C
from utils import *

class DeTextDataset(Dataset):
  def __init__(self, img_paths=None, mask_paths=None, poly_paths=None, a=True, test=False):
    self.a = a
    self.img_paths = img_paths
    self.mask_paths = mask_paths
    self.poly_paths = poly_paths
    assert len(self.img_paths) == len(self.mask_paths)
    self.images = len(self.img_paths) #list all the files present in that folder...
    self.test = test

  def __len__(self):
    return len(self.img_paths) #length of dataset

  def Lowpass(self, img):
    temp = img.copy()
    dst = cv2.GaussianBlur(temp,(5,5),cv2.BORDER_DEFAULT)
    return dst

  def Highpass(self, img):
    temp = img.copy()
    dst = cv2.GaussianBlur(temp, (3,3) ,0)
    source_gray = cv2.cvtColor(dst, cv2.COLOR_BGR2GRAY)
    dest = cv2.Laplacian(source_gray, cv2.CV_16S, ksize=3)
    abs_dest = cv2.convertScaleAbs(dest)
    return abs_dest

  def __getitem__(self, index):
    img_path = self.img_paths[index]
    mask_path = self.mask_paths[index]
    poly_path = self.poly_paths[index]

    image = cv2.imread(img_path)
    image = cv2.cvtColor(image, cv2.COLOR_BGR2RGB)
    image = cv2.resize(image, (512, 512))

    mask = cv2.imread(mask_path)
    mask = cv2.cvtColor(mask, cv2.COLOR_BGR2RGB)
    mask = cv2.resize(mask, (512, 512))

    poly = poly_to_mask(poly_path)
    if self.a:
      image, mask, poly = transforms(image, mask, poly)

    highpass = self.Highpass(mask)

    image = image.astype(np.float32)
    image = image/255.0
    image = torch.from_numpy(image)
    image = image.permute(2,0,1)

    poly = poly.astype('float32')
    poly = torch.from_numpy(poly)
    poly = (poly.permute(2, 0, 1) / 255.0)


    highpass = highpass.astype(np.float32)
    highpass = highpass[:,:,np.newaxis]
    highpass = highpass/255.0
    highpass = torch.from_numpy(highpass)
    highpass = highpass.permute(2,0,1)

    mask = mask.astype(np.float32)
    mask = mask/255.0
    mask = torch.from_numpy(mask)
    mask = mask.permute(2,0,1)

    return image, poly[0:1, :, :], highpass, mask


class TestDeTextDataset(Dataset):
  def __init__(self, img_paths=None, mask_paths=None, size=1024):

    self.img_paths = img_paths
    self.mask_paths = mask_paths
    self.size = size
    assert len(self.img_paths) == len(self.mask_paths)


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

  def __getitem__(self, index):
    img_path = self.img_paths[index]
    mask_path = self.mask_paths[index]

    image = cv2.imread(img_path)
    image = cv2.cvtColor(image, cv2.COLOR_BGR2RGB)
    image = cv2.resize(image, (self.size, self.size))

    image = image.astype(np.float32)
    image = image/255.0
    image = torch.from_numpy(image)
    image = image.permute(2,0,1)

    mask = cv2.imread(mask_path)
    mask = cv2.cvtColor(mask, cv2.COLOR_BGR2RGB)
    mask = cv2.resize(mask, (self.size, self.size))

    mask = mask.astype(np.float32)
    mask = mask/255.0
    mask = torch.from_numpy(mask)
    mask = mask.permute(2,0,1)

    return image, mask


def get_loaders():

    x_train=glob.glob(C.train_x)
    y_train=glob.glob(C.train_y)
    mask_trian=glob.glob(C.train_mask)

    x_test=glob.glob(C.test_x)
    y_test=glob.glob(C.test_y)
    mask_test=glob.glob(C.mask_test)

    train_ds = DeTextDataset(x_train, y_train,
    mask_train,
    a=True, test=False)

    test_ds = DeTextDataset(x_test, y_test,
    mask_test,
    a=False, test=True)


    train_loader = DataLoader(train_ds, batch_size=C.batch_size, num_workers=C.num_worker, shuffle=True)
    test_loader = DataLoader(test_ds, batch_size=C.batch_size, num_workers=C.num_worker, shuffle=False)

    return train_loader, test_loader