-
Notifications
You must be signed in to change notification settings - Fork 8
/
evaluation.py
190 lines (165 loc) · 7.81 KB
/
evaluation.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
import argparse
from datetime import datetime
import numpy as np
import tensorflow as tf
import socket
import importlib
import os
import sys
import scipy.io as sio
import copy
from sklearn.neighbors import NearestNeighbors
from scipy.misc import comb
import data_prep
import seg_util
parser = argparse.ArgumentParser()
parser.add_argument('--gpu', type=int, default=0, help='GPU to use [default: GPU 0]')
parser.add_argument('--model', default='model', help='Model name [default: model]')
parser.add_argument('--model_path', default='log3/best_model.ckpt', help='model checkpoint file path [default: log3/best_model.ckpt]')
FLAGS = parser.parse_args()
GPU_INDEX = FLAGS.gpu
NPOINT = 512
NMASK = 10
MODEL_DIR = os.path.dirname(os.path.abspath(__file__))
MODEL = importlib.import_module(FLAGS.model) # import network module
MODEL_FILE = os.path.join(MODEL_DIR, FLAGS.model+'.py')
MODEL_PATH = FLAGS.model_path
HOSTNAME = socket.gethostname()
TEST_DATASET = data_prep.SynTestDataset('data/sf2f_test.mat', npoint=NPOINT)
ISFULL_MATCHING = True
# TEST_DATASET = data_prep.SynTestDataset('data/sf2p_test.mat', npoint=NPOINT)
# ISFULL_MATCHING = False
def evaluation():
with tf.Graph().as_default():
with tf.device('/gpu:'+str(GPU_INDEX)):
pcpair_pl, flow_pl, _, _ = MODEL.placeholder_inputs(NMASK, NPOINT)
is_training_pl = tf.placeholder(tf.bool, shape=())
print("--- Get model and loss")
pred_flow, _ = MODEL.eva_flow(pcpair_pl)
pred_trans, pred_grouping, pred_seg, pred_conf = MODEL.eva_seg(pcpair_pl, flow_pl, nmask=NMASK)
saver = tf.train.Saver()
# Create a session
config = tf.ConfigProto()
config.gpu_options.allow_growth = True
config.allow_soft_placement = True
sess = tf.Session(config=config)
# Add summary writers
merged = tf.summary.merge_all()
# Init variables
init = tf.global_variables_initializer()
sess.run(init)
saver.restore(sess, MODEL_PATH)
ops = {'pcpair_pl': pcpair_pl,
'flow_pl': flow_pl,
'is_training_pl': is_training_pl,
'pred_flow': pred_flow,
'pred_trans': pred_trans,
'pred_grouping': pred_grouping,
'pred_seg': pred_seg,
'pred_conf': pred_conf,
'merged': merged}
eval(sess, ops)
def get_batch_data(dataset, idxs, start_idx, end_idx):
bsize = end_idx-start_idx
batch_pcpair = np.zeros((bsize, NPOINT, 6), dtype=np.float)
batch_flow = np.zeros((bsize, NPOINT, 3), dtype=np.float)
batch_seg = np.zeros((bsize, NPOINT), dtype=np.int)
for i in range(bsize):
pc1, pc2, flow12, seg1 = dataset[idxs[i+start_idx]]
batch_pcpair[i,:,:] = np.concatenate((pc1,pc2),1)
batch_flow[i,:,:] = flow12
batch_seg[i,:] = seg1
return batch_pcpair, batch_flow, batch_seg
def rand_index_score(label_pred, label_gt):
tp_plus_fp = comb(np.bincount(label_pred), 2).sum()
tp_plus_fn = comb(np.bincount(label_gt), 2).sum()
A = np.c_[(label_pred, label_gt)]
tp = sum(comb(np.bincount(A[A[:, 0] == i, 1]), 2).sum()
for i in set(label_pred))
fp = tp_plus_fp - tp
fn = tp_plus_fn - tp
tn = comb(len(A), 2) - tp - fp - fn
return (tp + tn) / (tp + fp + fn + tn)
def eval(sess, ops):
is_training = False
test_idxs = np.arange(0, len(TEST_DATASET))
num_ins = np.int(len(TEST_DATASET))
RI_prev_all = list()
RI_all = list()
EPE_all = list()
for ins_idx in range(0,num_ins,5):
print(ins_idx)
start_idx = ins_idx
end_idx = ins_idx+1
batch_pcpair, batch_flow, batch_seg = get_batch_data(TEST_DATASET, test_idxs, start_idx, end_idx)
pc1_ini = copy.deepcopy(batch_pcpair[0,:,:3])
pc1 = batch_pcpair[0,:,:3]
pc2 = batch_pcpair[0,:,3:6]
nbrs = NearestNeighbors(n_neighbors=2, algorithm='ball_tree').fit(pc1)
dd, idx = nbrs.kneighbors(pc1)
dist_th = np.max(dd[:,1])
nbrs = NearestNeighbors(n_neighbors=2, algorithm='ball_tree').fit(pc2)
dd, idx = nbrs.kneighbors(pc2)
dist_th = np.maximum(dist_th, np.max(dd[:,1]))
dist_th = np.minimum(dist_th, 0.06)
#### step1 fit global motion
feed_dict = {ops['pcpair_pl']: batch_pcpair,
ops['is_training_pl']: is_training}
flow_pred = sess.run(ops['pred_flow'], feed_dict=feed_dict)
pc1 = batch_pcpair[0,:,:3]
pc_pred = pc1+flow_pred[0,...]
## globally align everything
R, t = seg_util.fit_motion(pc1, pc_pred)
batch_pcpair[0,:,:3] = np.matmul(pc1,R)+t
feed_dict = {ops['pcpair_pl']: batch_pcpair,
ops['is_training_pl']: is_training}
flow_pred = sess.run(ops['pred_flow'], feed_dict=feed_dict)
pc1 = batch_pcpair[0,:,:3]
pc_pred = pc1+flow_pred[0,...]
## globally align everything
R, t = seg_util.fit_motion(pc1, pc_pred)
batch_pcpair[0,:,:3] = np.matmul(pc1,R)+t
feed_dict = {ops['pcpair_pl']: batch_pcpair,
ops['is_training_pl']: is_training}
flow_pred = sess.run(ops['pred_flow'], feed_dict=feed_dict)
pc1 = batch_pcpair[0,:,:3]
pc_pred = pc1+flow_pred[0,...]
#### iterative corrs and seg
expand_eps_list = [0.2]+[0.15]*4+[0.1]*4
for i, expand_eps in enumerate(expand_eps_list):
## seg
batch_pcpair = np.concatenate((np.expand_dims(pc1,0),np.expand_dims(pc2,0)),2)
feed_dict = {ops['pcpair_pl']: batch_pcpair,
ops['flow_pl']: np.expand_dims(pc_pred-pc1,0),
ops['is_training_pl']: is_training}
trans_pred, grouping_pred, seg_pred, conf_pred = sess.run([
ops['pred_trans'], ops['pred_grouping'], ops['pred_seg'], ops['pred_conf']], feed_dict=feed_dict)
Rmodes, tmodes, nmodes, segidx, segidx2 = seg_util.decode_motion_modes(pc1, pc_pred-pc1, pc2, trans_pred[0], grouping_pred[0], seg_pred[0], conf_pred[0], eps=dist_th)
subpc1, subpc2, attmask1, attmask2, segidx, segidx2, distmatrix, distmatrix2 = seg_util.gen_attention_mask(pc1, pc_pred, pc2, segidx, segidx2, Rmodes, tmodes, nmodes, NPOINT, expand_eps=expand_eps)
for j in range(nmodes):
subpc1[j,:,:] = np.matmul(subpc1[j,:,:], Rmodes[j,:,:])+tmodes[[j],:]
## corrs
batch_pcpair = np.concatenate((subpc1, subpc2),-1)
feed_dict = {ops['pcpair_pl']: batch_pcpair,
ops['is_training_pl']: is_training}
flow_pred = sess.run(ops['pred_flow'], feed_dict=feed_dict)
pc_pred, segidx = seg_util.motion_modes_aggreg_watt(subpc1, subpc1+flow_pred, attmask1)
if ISFULL_MATCHING:
pc_pred = pc2[np.argmin(np.sum((np.expand_dims(pc_pred,1)-np.expand_dims(pc2,0))**2,2),1),:]
batch_pcpair = np.concatenate((np.expand_dims(pc1,0),np.expand_dims(pc2,0)),2)
feed_dict = {ops['pcpair_pl']: batch_pcpair,
ops['flow_pl']: np.expand_dims(pc_pred-pc1,0),
ops['is_training_pl']: is_training}
trans_pred, grouping_pred, seg_pred, conf_pred = sess.run([
ops['pred_trans'], ops['pred_grouping'], ops['pred_seg'], ops['pred_conf']], feed_dict=feed_dict)
## final seg
seg_gt = batch_seg[0].reshape(-1)
seg_pred = np.argmax(seg_pred[0][np.squeeze(conf_pred[0])>0.5,:],0).reshape(-1)
seg_pred = seg_util.seg_merge(pc1, pc_pred, seg_pred)
RI_all.append(rand_index_score(seg_pred, seg_gt))
EPE_all.append( np.mean(np.sqrt(np.sum((pc_pred-pc1_ini-batch_flow[0])**2,1))) )
print('Mean RI: %f'%np.mean(np.array(RI_all)))
print('Mean EPE: %f'%np.mean(np.array(EPE_all)))
return 0
if __name__ == "__main__":
evaluation()