-
Notifications
You must be signed in to change notification settings - Fork 9
/
dataset.py
70 lines (53 loc) · 2.09 KB
/
dataset.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
import cv2
import torch
from torch.utils.data.dataset import Dataset
import random as random
import numpy as np
import matplotlib.pyplot as plt
class CustomDataset(Dataset):
def __init__(self, args):
self.args = args
height, width = args.image_size, args.image_size
self.data = []
self.index = []
print(f'Generating samples...')
for i in range(args.dataset_size):
x, y, t = 0, 0, 0
c = 0
offset = random.randint(1, int(height/2))
seq_idx = {
"idx_from": len(self.data),
"idx_to": 0
}
while True:
t += 1
img = np.zeros((height, width, 1), np.float32)
img.fill(1)
img = cv2.circle(img, (int(x), int(abs(y)) + args.r + offset), args.r, (0), -1)
self.data.append(img)
x = args.vx * t
y = args.vy * t - 1 / 2 * args.g * t ** 2
if abs(y) > height - args.r - offset:
break
c += 1
seq_idx["idx_to"] = len(self.data)
self.index.append(seq_idx)
# self.plot()
def plot(self):
i = self.index[0]
d = self.data[i['idx_from']: i['idx_to']]
f, axarr = plt.subplots(1, len(d))
for plot, img in enumerate(d):
axarr[plot].imshow(np.squeeze(img, -1), cmap='gray')
plt.show()
def __getitem__(self, idx):
i = self.index[idx]
imgs = self.data[i["idx_from"] : i["idx_from"] + self.args.sequence_window]
imgs_truths = self.data[i["idx_from"] + self.args.sequence_window : i["idx_to"]]
imgs = torch.FloatTensor(imgs)
imgs = imgs.permute(0, 3, 1, 2) # (B, H, W, C) --> (B, C, H, W)
imgs_truths = torch.FloatTensor(imgs_truths)
imgs_truths = imgs_truths.permute(0, 3, 1, 2) # (B, H, W, C) --> (B, C, H, W)
return {'imgs': imgs, 'imgs_truths': imgs_truths }
def __len__(self):
return len(self.index)