-
Notifications
You must be signed in to change notification settings - Fork 0
/
Copy pathcal_metric_2.py
136 lines (107 loc) · 6.38 KB
/
cal_metric_2.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
import numpy as np
import open3d as o3d
import sklearn.neighbors as skln
import argparse
from glob import glob
import os
if __name__ == "__main__":
parser = argparse.ArgumentParser()
parser.add_argument('--data_dir', type=str, default='/data/huyb/nips-2023/ys-NeuRIS/exps-paper/indoor-paper/collected_meshes_per_scene/newfilled_5', help='input data')
parser.add_argument('--gt_dir', type=str, default='/data/huyb/nips-2023/ys-NeuRIS/evaluation/meshes-scan2cad', help='gt data')
parser.add_argument('--vis_out_dir', type=str, default='/data/huyb/nips-2023/ys-NeuRIS/exps-paper/indoor-paper/collected_meshes_per_scene/newfilled_5/metrics')
parser.add_argument('--max_dist', type=float, default=10, help='ignore the irrelevant points')
args = parser.parse_args()
# for scene_name in ['scene0008_00', 'scene0005_00', 'scene0050_00', 'scene0461_00', 'scene0549_00', 'scene0616_00']:
for scene_name in ['scene0616_00']:
scene_dir = os.path.join(args.data_dir, scene_name)
# gt_dir = os.path.join(args.gt_dir, scene_name[5:9], 'gt')
gt_dir = os.path.join(args.gt_dir, scene_name)
input_data_list = sorted(glob(os.path.join(scene_dir, '*_full_mesh.ply')))
if not os.path.exists(args.vis_out_dir):
os.mkdir(args.vis_out_dir)
file = open(os.path.join(args.vis_out_dir, f"metric_{scene_name}_scan2cad.txt"), 'w')
accuracy_list = []
completion_list = []
chamfer_distance_list = []
precision_1_list = []
recall_1_list = []
fscore_1_list = []
fscore_2_list = []
fscore_3_list = []
for input_data in input_data_list:
obj_id = os.path.basename(input_data).split('_')[4]
if scene_name != 'scene0616_00':
obj_id = int(obj_id) - 1
print(obj_id)
gt_data = os.path.join(gt_dir, f"scannet_{obj_id}_pointcloud.ply")
if not os.path.exists(gt_data):
continue
print(input_data)
print(gt_data)
dist_thred3 = 0.10
dist_thred1 = 0.05 # 5cm
dist_thred2 = 0.02 # 2cm
print('read data pcd ...')
data_pcd_o3d = o3d.io.read_point_cloud(input_data)
data_pcd = np.asarray(data_pcd_o3d.points)
print('random shuffle pcd index ...')
shuffle_rng = np.random.default_rng()
shuffle_rng.shuffle(data_pcd, axis=0)
print('read GT pcd ...')
stl_pcd = o3d.io.read_point_cloud(gt_data)
stl = np.asarray(stl_pcd.points)
print('compute data2gt ...')
nn_engine = skln.NearestNeighbors(n_neighbors=1, radius=1.0, algorithm='kd_tree', n_jobs=-1)
nn_engine.fit(stl)
dist_d2s, idx_d2s = nn_engine.kneighbors(data_pcd, n_neighbors=1, return_distance=True)
max_dist = args.max_dist
mean_d2s = dist_d2s[dist_d2s < max_dist].mean()
precision_1 = len(dist_d2s[dist_d2s < dist_thred1]) / len(dist_d2s)
precision_2 = len(dist_d2s[dist_d2s < dist_thred2]) / len(dist_d2s)
precision_3 = len(dist_d2s[dist_d2s < dist_thred3]) / len(dist_d2s)
print('compute gt2data ...')
nn_engine.fit(data_pcd)
dist_s2d, idx_s2d = nn_engine.kneighbors(stl, n_neighbors=1, return_distance=True)
mean_s2d = dist_s2d[dist_s2d < max_dist].mean()
recall_1 = len(dist_s2d[dist_s2d < dist_thred1]) / len(dist_s2d)
recall_2 = len(dist_s2d[dist_s2d < dist_thred2]) / len(dist_s2d)
recall_3 = len(dist_s2d[dist_s2d < dist_thred3]) / len(dist_s2d)
over_all = (mean_d2s + mean_s2d) / 2
fscore_1 = 2 * precision_1 * recall_1 / (precision_1 + recall_1 + 1e-8)
fscore_2 = 2 * precision_2 * recall_2 / (precision_2 + recall_2 + 1e-8)
fscore_3 = 2 * precision_3 * recall_3 / (precision_3 + recall_3 + 1e-8)
accuracy_list.append(mean_d2s)
completion_list.append(mean_s2d)
chamfer_distance_list.append(over_all)
fscore_1_list.append(fscore_1)
precision_1_list.append(precision_1)
recall_1_list.append(recall_1)
fscore_2_list.append(fscore_2)
fscore_3_list.append(fscore_3)
print(f'chamfer_distance_1m: {over_all}; mean_d2gt_1m: {mean_d2s}; mean_gt2d_1m: {mean_s2d}.')
print(f'precision_10cm: {precision_3}; recall_10cm: {recall_3}; fscore_10cm: {fscore_3}')
print(f'precision_5cm: {precision_1}; recall_5cm: {recall_1}; fscore_5cm: {fscore_1}')
print(f'precision_2cm: {precision_2}; recall_2cm: {recall_2}; fscore_2cm: {fscore_2}')
file.write(f'{obj_id}' + '\n')
file.write(f'chamfer_distance_1m: {over_all}; mean_d2gt_1m: {mean_d2s}; mean_gt2d_1m: {mean_s2d}.' + '\n')
file.write(f'precision_10cm: {precision_3}; recall_10cm: {recall_3}; fscore_10cm: {fscore_3}' + '\n')
file.write(f'precision_5cm: {precision_1}; recall_5cm: {recall_1}; fscore_5cm: {fscore_1}' + '\n')
file.write(f'precision_2cm: {precision_2}; recall_2cm: {recall_2}; fscore_2cm: {fscore_2}' + '\n')
file.write('\n')
accuracy_avg = sum(accuracy_list) / len(accuracy_list)
completion_avg = sum(completion_list) / len(completion_list)
chamfer_distance_avg = sum(chamfer_distance_list) / len(chamfer_distance_list)
f_score1_avg = sum(fscore_1_list) / len(fscore_1_list)
precision_1_avg = sum(precision_1_list) / len(precision_1_list)
recall_1_avg = sum(recall_1_list) / len(recall_1_list)
f_score2_avg = sum(fscore_2_list) / len(fscore_2_list)
f_score3_avg = sum(fscore_3_list) / len(fscore_3_list)
file.write(f'accuracy_avg (1m): {accuracy_avg};' + '\n')
file.write(f'completion_avg (1m): {completion_avg};' + '\n')
file.write(f'chamfer_distance_avg (1m): {chamfer_distance_avg};' + '\n')
file.write(f'fscore_3_avg (10cm): {f_score3_avg};' + '\n')
file.write(f'fscore_1_avg (5cm): {f_score1_avg};' + '\n')
file.write(f'precision_1_avg (5cm): {precision_1_avg};' + '\n')
file.write(f'recall_1_avg (5cm): {recall_1_avg};' + '\n')
file.write(f'fscore_2_avg (2cm): {f_score2_avg};' + '\n')
file.close()