-
Notifications
You must be signed in to change notification settings - Fork 0
/
Copy pathMain3.py
211 lines (163 loc) · 8.81 KB
/
Main3.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
147
148
149
150
151
152
153
154
155
156
157
158
159
160
161
162
163
164
165
166
167
168
169
170
171
172
173
174
175
176
177
178
179
180
181
182
183
184
185
186
187
188
189
190
191
192
193
194
195
196
197
198
199
200
201
202
203
204
205
206
207
208
209
210
211
import open3d as open3d
import networkx as nx
import numpy as np
import scipy.sparse as spsp
import scipy.signal as spsig
import scipy as sp
import scipy.cluster.vq as vq
import spectral_clustering.similarity_graph as SGk
import spectral_clustering.branching_graph as BGk
import sklearn.cluster as sk
from sklearn.metrics.pairwise import euclidean_distances
import matplotlib.pyplot as plt
## A partir de l'import d'un fichier texte
# Calcul des Kmeans sur un nuage de points en fichier .txt
# En prenant en compte les coordonnées et le label.
#pointcloud_with_label = np.loadtxt(fname="/Users/katiamirande/Documents/Tests/Tests graphes simples/Données_hypothèse_monotonie/Branches_t20_espacement_aleatoire/testchain1.txt"
# , delimiter=',', usecols=range(4))
#pointcloud_with_label_sort = pointcloud_with_label[pointcloud_with_label[:,3].argsort()]
#vp2grad = np.gradient(pointcloud_with_label_sort[:,3])
"""
G = BGk.BranchingGraph(500)
G.add_branch(branch_size=100, linking_node=200)
G.add_branch(branch_size=20, linking_node=530)
G.add_branch(branch_size=10, linking_node=610)
G.add_branch(branch_size=100, linking_node=350)
G.add_branch(branch_size=20, linking_node=665)
G.add_branch(branch_size=10, linking_node=745)
G.compute_graph_eigenvectors()
#labels = G.keigenvec[:, 1]
#labels_sorts = labels[labels.argsort()]
#vp2grad = np.gradient(labels_sorts)
vp2grad = np.gradient(labels)
figureval = plt.figure(0)
figureval.clf()
figureval.gca().plot(range(len(np.transpose(vp2grad))), np.transpose(vp2grad), 'bo')
figureval.set_size_inches(20, 10)
figureval.subplots_adjust(wspace=0, hspace=0)
figureval.tight_layout()
figureval.savefig('/Users/katiamirande/Documents/Tests/Tests_gradient__sort_vp2')
"""
#c = 2
#means,labels = vq.kmeans2(pointcloud_with_label, c, minit='points', missing='warn')
#pcdtabclassif = sk.agglomerativeClustering().fit(pointcloud_with_label)
#labels = np.asarray(labels.reshape(pointcloud_with_label.shape[0], 1), dtype= np.float64)
#np.savetxt('Centroides.txt', means, delimiter= ',')
#np.savetxt('/Users/katiamirande/Documents/Tests/Tests graphes simples/Données/Chaine100Branchet25_n35/PCD_withvp2_segmented_with_agglomerativeclustering.txt', pcdtabclassif, delimiter = ",")
## Création du graph pour obtenir les données associées.
# Objectif : utilisation d'un algorithme sci kit : agglomerative clustering.
"""
#G = BGk.BranchingGraph(100)
#G.add_branch(branch_size=20, linking_node=35)
#G = BGk.BranchingGraph(100)
#G.add_branch(branch_size=20, linking_node=30)
#G.add_branch(branch_size=20, linking_node=35)
#G.add_branch(branch_size=20, linking_node=40)
#G.add_branch(branch_size=20, linking_node=47)
#G.add_branch(branch_size=20, linking_node=50)
#G.add_branch(branch_size=20, linking_node=55)
#G.add_branch(branch_size=20, linking_node=62)
#G.add_branch(branch_size=20, linking_node=63)
#G.add_branch(branch_size=20, linking_node=70)
G = BGk.BranchingGraph(500)
G.add_branch(branch_size=100, linking_node=200)
G.add_branch(branch_size=20, linking_node=530)
G.add_branch(branch_size=10, linking_node=610)
def clustering_by_fiedler_and_agglomerative(self):
G.compute_graph_eigenvectors()
G.add_eigenvector_value_as_attribute(k=2)
A = nx.adjacency_matrix(G)
vp2 = np.asarray(G.keigenvec[:,1])
vp2_matrix = np.tile(vp2,(len(G),1))
vp2_matrix[A.todense()==0] = np.nan
vp2grad = (np.nanmax(vp2_matrix,axis=1) - np.nanmin(vp2_matrix,axis=1))/2.
node_vp2grad_values = dict(zip(G.nodes(), vp2grad))
nx.set_node_attributes(G, node_vp2grad_values, 'gradient_vp2')
clustering = sk.AgglomerativeClustering(affinity='euclidean', connectivity=A, linkage='single', n_clusters=6).fit(vp2grad[:, np.newaxis])
node_clustering_label = dict(zip(G.nodes(), np.transpose(clustering.labels_)))
nx.set_node_attributes(G, node_clustering_label, 'clustering_label')
BGk.save_graph_plot(G, ['gradient_vp2'], colormap='Reds',
filename='/Users/katiamirande/Documents/Tests/Tests_gradient_vp2/Tests_gradientminmax_agglomerativeclustering/valeurs_gradient_vp2')
BGk.save_graph_plot(G, ['clustering_label'], colormap='plasma', filename='/Users/katiamirande/Documents/Tests/Tests_gradient_vp2/Tests_gradientminmax_agglomerativeclustering/clusters_gradient_vp2')
"""
pcd = open3d.read_point_cloud("Data/arabi_densep_clean_segm.ply")
r = 8
G = SGk.create_riemannian_graph(pcd, method='knn', nearest_neighbors=r)
Lcsr = nx.laplacian_matrix(G, weight='weight')
Lcsr = spsp.csr_matrix.asfptype(Lcsr)
# k nombre de vecteurs propres que l'on veut calculer
k = 2
keigenval, keigenvec = spsp.linalg.eigsh(Lcsr,k=k,sigma=0, which='LM')
eigenval = keigenval.reshape(keigenval.shape[0], 1)
A = nx.adjacency_matrix(G)
vp2 = np.asarray(keigenvec[:, 1])
pcd_vp2 = np.concatenate([np.asarray(pcd.points), vp2[:,np.newaxis]], axis=1)
np.savetxt('pcd_vp2.txt', pcd_vp2, delimiter=",")
print("Export du nuage avec vecteur propre 2")
vp2_matrix = A.copy()
vp2_matrix[A.nonzero()] = vp2[A.nonzero()[1]]
#Second method fo gradient, just obtain the max and min value in the neighborhood.
node_neighbor_max_vp2 = np.array([vp2_matrix[node].data.max() for node in range(A.shape[0])])
node_neighbor_min_vp2 = np.array([vp2_matrix[node].data.min() for node in range(A.shape[0])])
vp2grad = node_neighbor_max_vp2 - node_neighbor_min_vp2
#First method not adapted to big clouds
#node_neighbor_max_vp2 = np.array([vp2[A[node].nonzero()[1]].max() for node in range(A.shape[0])])
#node_neighbor_min_vp2 = np.array([vp2[A[node].nonzero()[1]].min() for node in range(A.shape[0])])
"""
node_neighbor_max_vp2_node = np.array([np.argmax(vp2_matrix[node].data) for node in range(A.shape[0])])
node_neighbor_max_vp2 = np.array([vp2_matrix[node].data[neighbor_max_node] for node,neighbor_max_node in zip(range(A.shape[0]),node_neighbor_max_vp2_node)])
node_neighbor_min_vp2_node = np.array([np.argmin(vp2_matrix[node].data) for node in range(A.shape[0])])
node_neighbor_min_vp2 = np.array([vp2_matrix[node].data[neighbor_min_node] for node,neighbor_min_node in zip(range(A.shape[0]),node_neighbor_min_vp2_node)])
pcdtab = np.asarray(pcd.points)
vp2_max_min = (node_neighbor_max_vp2 - node_neighbor_min_vp2)
grad_weight = np.zeros(node_neighbor_min_vp2_node.shape[0])
for i in range(node_neighbor_min_vp2_node.shape[0]):
grad_weight[i] = sp.spatial.distance.euclidean(pcdtab[node_neighbor_max_vp2_node[i]].reshape(1, 3), pcdtab[node_neighbor_min_vp2_node[i]].reshape(1, 3))
vp2grad = np.divide(vp2_max_min, grad_weight)
"""
pcd_vp2_grad = np.concatenate([np.asarray(pcd.points), vp2grad[:,np.newaxis]], axis=1)
np.savetxt('pcd_vp2_grad.txt', pcd_vp2_grad, delimiter=",")
print("Export du nuage avec gradient du vecteur propre 2")
pcd_vp2_grad_vp2 = np.concatenate([pcd_vp2_grad, vp2[:,np.newaxis]], axis=1)
pcd_vp2_grad_vp2_sort_by_vp2 = pcd_vp2_grad_vp2[pcd_vp2_grad_vp2[:,4].argsort()]
figure = plt.figure(0)
figure.clf()
figure.gca().set_title("fiedler vector")
# figure.gca().plot(range(len(vec)),vec,color='blue')
figure.gca().scatter(range(len(pcd_vp2_grad_vp2_sort_by_vp2)), pcd_vp2_grad_vp2_sort_by_vp2[:, 4], color='blue')
figure.set_size_inches(10, 10)
figure.subplots_adjust(wspace=0, hspace=0)
figure.tight_layout()
figure.savefig("fiedler_vector")
print("Export du vecteur propre 2")
figure = plt.figure(1)
figure.clf()
figure.gca().set_title("Gradient of fiedler vector")
# figure.gca().plot(range(len(vec)),vec,color='blue')
figure.gca().scatter(range(len(pcd_vp2_grad_vp2_sort_by_vp2)), pcd_vp2_grad_vp2_sort_by_vp2[:, 3], color='blue')
figure.set_size_inches(10, 10)
figure.subplots_adjust(wspace=0, hspace=0)
figure.tight_layout()
figure.savefig("Gradient_of_fiedler_vector")
print("Export du gradient trié selon vecteur propre 2")
# vp2_matrix = np.tile(vp2, (len(G), 1))
# vp2_matrix[A.todense() == 0] = np.nan
# vp2grad = (np.nanmax(vp2_matrix, axis=1) - np.nanmin(vp2_matrix, axis=1)) / 2.
node_vp2grad_values = dict(zip(G.nodes(), vp2grad))
#mean_gradient = np.mean(vp2grad)
#maximas_grad = np.diff(np.sign(np.diff(vp2grad)))
#number_maximas = np.count_nonzero(maximas_grad)
#number_maximas = sum(1 for i in maximas_grad if i < 0)
#clustering via nombre de clusters
clustering = sk.AgglomerativeClustering(affinity='euclidean', connectivity=A, linkage='ward', n_clusters=11).fit(vp2grad[:, np.newaxis])
#clustering via distance_threshold
#clustering = sk.AgglomerativeClustering(affinity='euclidean',
# connectivity=A,
# linkage='ward',
# n_clusters=None,
# compute_full_tree=True,
# distance_threshold=mean_gradient).fit(vp2grad[:, np.newaxis])
label = np.asarray(clustering.labels_)
pcdtabclassif = np.concatenate([np.asarray(pcd.points), label[:,np.newaxis]], axis=1)
np.savetxt('pcdclassifagglo11.txt', pcdtabclassif, delimiter=",")
print("Export du nuage classifié")