UnlearnablePC/data_utils/ScanObjectNNDataLoader.py at main · CGCL-codes/UnlearnablePC · 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
137
138
139
140
import numpy as np
import h5py
import os
from torch.utils.data import Dataset


def load_h5(h5_filename):
    f = h5py.File(h5_filename, 'r')
    data = f['data'][:]
    label = f['label'][:]
    return data, label

def get_current_data_h5(pcs, labels, num_points):
	#shuffle points to sample
	idx_pts = np.arange(pcs.shape[1])
	np.random.shuffle(idx_pts)

	sampled = pcs[:,idx_pts[:num_points],:]
	#sampled = pcs[:,:num_points,:]

	#shuffle point clouds per epoch
	idx = np.arange(len(labels))
	np.random.shuffle(idx)

	sampled = sampled[idx]
	labels = labels[idx]

	return sampled, labels


def pc_normalize(pc):
    centroid = np.mean(pc, axis=0)
    pc = pc - centroid
    m = np.max(np.sqrt(np.sum(pc**2, axis=1)))
    pc = pc / m
    return pc


def farthest_point_sample(point, npoint):
    """
    Input:
        xyz: pointcloud data, [N, D]
        npoint: number of samples
    Return:
        centroids: sampled pointcloud index, [npoint, D]
    """
    N, D = point.shape
    xyz = point[:,:3]
    centroids = np.zeros((npoint,))
    distance = np.ones((N,)) * 1e10
    farthest = np.random.randint(0, N)
    for i in range(npoint):
        centroids[i] = farthest
        centroid = xyz[farthest, :]
        dist = np.sum((xyz - centroid) ** 2, -1)
        mask = dist < distance
        distance[mask] = dist[mask]
        farthest = np.argmax(distance, -1)
    point = point[centroids.astype(np.int32)]
    return point


class ScanObjectNNDataLoader(Dataset):
    def __init__(self, root,  npoint=1024, split='train', train_num = 1000, uniform=False, cache_size=15000):
        self.root = root
        self.npoint = npoint
        self.uniform = uniform

        #! random choice 2662
        assert (split == 'train' or split == 'test')
        if split == 'train':
            self.data, self.labels = load_h5(os.path.join(self.root, 'main_split_nobg/training_objectdataset.h5'))
        else:
            self.data, self.labels = load_h5(os.path.join(self.root, 'main_split_nobg/test_objectdataset.h5'))

        print('The size of %s data is %d'%(split,len(self.data)))
        self.cache_size = cache_size  # how many data points to cache in memory
        self.cache = {}  # from index to (point_set, cls) tuple


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

    def _get_item(self, index):
        if index in self.cache:
            point_set, cls = self.cache[index]
        else:
            point_set = self.data[index]
            cls = self.labels[index]
            cls = np.array([cls]).astype(np.int32)
            point_set = point_set.astype(np.float32)
            if self.uniform:
                point_set = farthest_point_sample(point_set, self.npoint)
            else:
                point_set = point_set[0:self.npoint,:]

            point_set[:, 0:3] = pc_normalize(point_set[:, 0:3])

            if len(self.cache) < self.cache_size:
                self.cache[index] = (point_set, cls)

        return point_set, cls

    def __getitem__(self, index):
        return self._get_item(index)

import  matplotlib.pyplot as plt
def plot_pcd_three_views(filename, pcds, titles, suptitle='', sizes=None, cmap='GnBu', zdir='y',
                        xlim=(-0.9, 0.9), ylim=(-0.9, 0.9), zlim=(-0.9, 0.9)):
    if sizes is None:
        sizes = [0.35 for _ in range(len(pcds))]
    fig = plt.figure(figsize=(len(pcds) * 3, 9))
    for i in range(3):
        elev = 30
        azim = -45 + 90 * i
        for j, (pcd, size) in enumerate(zip(pcds, sizes)):
            color = pcd[:, 0]
            ax = fig.add_subplot(3, len(pcds), i * len(pcds) + j + 1, projection='3d')
            ax.view_init(elev, azim)
            ax.scatter(pcd[:, 0], pcd[:, 1], pcd[:, 2], zdir=zdir, c=color, s=size, cmap=cmap, vmin=-1, vmax=0.5)
            ax.set_title(titles[i])
            ax.set_axis_off()
            ax.set_xlim(xlim)
            ax.set_ylim(ylim)
            ax.set_zlim(zlim)
    plt.subplots_adjust(left=0.05, right=0.95, bottom=0.05, top=0.9, wspace=0.1, hspace=0.1)
    plt.suptitle(suptitle)
    fig.savefig(filename)
    plt.close(fig)


if __name__ == '__main__':
    import torch
    # 15 classes
    data = ScanObjectNNDataLoader('data/h5_files/',split='train', uniform=False)  # 2309, 581
    DataLoader = torch.utils.data.DataLoader(data, batch_size=1, shuffle=False)
    for point,label in DataLoader:
        print(label)
        plot_pcd_three_views('ttt.png', [point[0]], ['a', 'b', 'c'])
        break