-
Notifications
You must be signed in to change notification settings - Fork 4
/
Copy pathdata.py
146 lines (120 loc) · 4.76 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
138
139
140
141
142
143
144
145
146
import os
import glob
import h5py
import numpy as np
from torch.utils.data import Dataset
import random
from PIL import Image
from .plyfile import load_ply
from . import data_utils as d_utils
import torchvision.transforms as transforms
from PIL import ImageFile
ImageFile.LOAD_TRUNCATED_IMAGES = True
trans_1 = transforms.Compose(
[
d_utils.PointcloudToTensor(),
d_utils.PointcloudNormalize(),
d_utils.PointcloudScale(lo=0.5, hi=2, p=1),
d_utils.PointcloudRotate(),
d_utils.PointcloudTranslate(0.5, p=1),
d_utils.PointcloudJitter(p=1),
d_utils.PointcloudRandomInputDropout(p=1),
])
trans_2 = transforms.Compose(
[
d_utils.PointcloudToTensor(),
d_utils.PointcloudNormalize(),
d_utils.PointcloudScale(lo=0.5, hi=2, p=1),
d_utils.PointcloudRotate(),
d_utils.PointcloudTranslate(0.5, p=1),
d_utils.PointcloudJitter(p=1),
d_utils.PointcloudRandomInputDropout(p=1),
])
def load_modelnet_data(partition):
# change to your own path
DATA_DIR = '/mnt/sdb/public/data/common-datasets'
# DATA_DIR = os.path.join(BASE_DIR, 'data')
all_data = []
all_label = []
for h5_name in glob.glob(os.path.join(DATA_DIR, 'modelnet40_ply_hdf5_2048', 'ply_data_%s*.h5'%partition)):
f = h5py.File(h5_name)
data = f['data'][:].astype('float32')
label = f['label'][:].astype('int64')
f.close()
all_data.append(data)
all_label.append(label)
all_data = np.concatenate(all_data, axis=0)
all_label = np.concatenate(all_label, axis=0)
return all_data, all_label
def load_ScanObjectNN(partition):
# change to your own path
BASE_DIR = '/mnt/sdb/public/data/common-datasets/ScanObjectNN'
DATA_DIR = os.path.join(BASE_DIR, 'main_split')
h5_name = os.path.join(DATA_DIR, f'{partition}.h5')
f = h5py.File(h5_name)
data = f['data'][:].astype('float32')
label = f['label'][:].astype('int64')
return data, label
def load_shapenet_data():
DATA_DIR = '/mnt/sdb/public/data/common-datasets'
# DATA_DIR = os.path.join(BASE_DIR, 'data')
all_filepath = []
for cls in glob.glob(os.path.join(DATA_DIR, 'ShapeNet/*')):
pcs = glob.glob(os.path.join(cls, '*'))
all_filepath += pcs
return all_filepath
def get_render_imgs(pcd_path):
path_lst = pcd_path.split('/')
# path_lst[1] = 'ShapeNetRendering'
path_lst[-3] = 'ShapeNetRendering'
path_lst[-1] = path_lst[-1][:-4]
path_lst.append('rendering')
DIR = '/'.join(path_lst)
img_path_list = glob.glob(os.path.join(DIR, '*.png'))
return img_path_list
class ShapeNetRender(Dataset):
def __init__(self, img_transform = None, n_imgs = 1):
self.data = load_shapenet_data()
self.transform = img_transform
self.n_imgs = n_imgs
def __getitem__(self, item):
pcd_path = self.data[item]
render_img_path = random.choice(get_render_imgs(pcd_path))
# render_img_path_list = random.sample(get_render_imgs(pcd_path), self.n_imgs)
# render_img_list = []
# for render_img_path in render_img_path_list:
render_img = Image.open(render_img_path).convert('RGB')
render_img = self.transform(render_img) #.permute(1, 2, 0)
# render_img_list.append(render_img)
pointcloud_1 = load_ply(self.data[item])
# pointcloud_orig = pointcloud_1.copy()
pointcloud_2 = load_ply(self.data[item])
point_t1 = trans_1(pointcloud_1)
point_t2 = trans_2(pointcloud_2)
# pointcloud = (pointcloud_orig, point_t1, point_t2)
pointcloud = (point_t1, point_t2)
return pointcloud, render_img # render_img_list
def __len__(self):
return len(self.data)
class ModelNet40SVM(Dataset):
def __init__(self, num_points, partition='train'):
self.data, self.label = load_modelnet_data(partition)
self.num_points = num_points
self.partition = partition
def __getitem__(self, item):
pointcloud = self.data[item][:self.num_points]
label = self.label[item]
return pointcloud, label
def __len__(self):
return self.data.shape[0]
class ScanObjectNNSVM(Dataset):
def __init__(self, num_points, partition='train'):
self.data, self.label = load_ScanObjectNN(partition)
self.num_points = num_points
self.partition = partition
def __getitem__(self, item):
pointcloud = self.data[item][:self.num_points]
label = self.label[item]
return pointcloud, label
def __len__(self):
return self.data.shape[0]