-
Notifications
You must be signed in to change notification settings - Fork 2
/
evaluate.py
140 lines (121 loc) · 4.55 KB
/
evaluate.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
import numpy as np
import os
import pickle as p
import argparse
parser = argparse.ArgumentParser()
parser.add_argument('--result_dir', type=str,required=True)
args = parser.parse_args()
def compute_iou(pred_box, ref_bbox):
N=pred_box.shape[0]
pred_box_rb = pred_box[:,0:3] - pred_box[:,3:6] / 2.0
ref_bbox_rb = ref_bbox[:,0:3] - ref_bbox[:,3:6] / 2.0
pred_box_lt = pred_box[:,0:3] + pred_box[:,3:6] / 2.0
ref_bbox_lt = ref_bbox[:,0:3] + ref_bbox[:,3:6] / 2.0
lt = np.min(np.concatenate([pred_box_lt[:,:,np.newaxis],np.repeat(ref_bbox_lt[:,:,np.newaxis],N,axis=0)],axis=2),axis=2)
rb = np.max(np.concatenate([pred_box_rb[:,:,np.newaxis],np.repeat(ref_bbox_rb[:,:,np.newaxis],N,axis=0)],axis=2),axis=2)
whz = lt - rb
whz[whz < 0] = 0
inter = whz[:, 0] * whz[:, 1] * whz[:, 2]
pred_box_area = pred_box[:, 3] * pred_box[:, 4] * pred_box[:, 5]
ref_box_area = ref_bbox[:, 3] * ref_bbox[:, 4] * ref_bbox[:, 5]
# print(pred_box_area.shape,inter.shape,ref_box_area.shape)
iou = inter / (pred_box_area + np.repeat(ref_box_area,N,axis=0) - inter)
# print(iou)
return iou
result_dir=args.result_dir
k=5
success_count_iou25=0
success_count_iou50=0
Rat2_count=0
Rat5_count=0
Rat10_count=0
Rat20_count=0
total_count=0
Max_IoU=0
success_count=0
iou_sum=0
par_success_count_iou25=0
par_success_count_iou50=0
par_Rat2_count=0
par_Rat5_count=0
par_Rat10_count=0
par_Rat20_count=0
par_Max_IoU=0
scan_list=os.listdir(result_dir)
#print(scan_list)
#scan_list=[scan_list[0]]
for scan_file in scan_list:
scan_output_file=os.path.join(result_dir,scan_file)
scan=scan_file[:12]
#print(scan)
with open(scan_output_file,"rb") as f:
output_content=p.load(f)
object_id_list=list(output_content.keys())
for object_id in object_id_list:
for object_data in output_content[object_id]:
prediction=object_data["pred_intact_box"].T
gt=object_data["gt_intact_bbox"].T
partial_pred=object_data["pred_partial_box"]
partial_gt=object_data["gt_partial_bbox"].T
#prediction=partial_pred[:,0:6]
#gt=partial_gt
#print(prediction.shape)
bbox=prediction[:,0:6]
#print(prediction.shape)
partial_bbox=partial_pred[:,0:6]
confidence=object_data["output"][:,6]
sort_id=np.argsort(-confidence)
topk_id=sort_id[:20]
topk_bbox=bbox[topk_id]
topk_partial_bbox=partial_bbox[topk_id]
target_bbox=gt[np.newaxis,:]
target_partial_bbox=partial_gt[np.newaxis,:]
#print(target_bbox.shape)
iou=compute_iou(topk_bbox,target_bbox)
par_iou=compute_iou(topk_partial_bbox,target_partial_bbox)
par_Max_IoU+=np.max(par_iou)
Max_IoU+=np.max(iou)
if iou[0]>0.25:
success_count_iou25+=1
if iou[0]>0.5:
success_count_iou50+=1
iou_sum+=iou[0]
success_count+=1
if np.max(iou[0:2])>0.5:
Rat2_count+=1
if np.max(iou[0:5])>0.5:
Rat5_count+=1
if np.max(iou[0:10])>0.5:
Rat10_count+=1
if np.max(iou[0:20])>0.5:
Rat20_count+=1
if par_iou[0]>0.25:
par_success_count_iou25+=1
if par_iou[0]>0.5:
par_success_count_iou50+=1
if np.max(par_iou[0:2])>0.5:
par_Rat2_count+=1
if np.max(par_iou[0:5])>0.5:
par_Rat5_count+=1
if np.max(par_iou[0:10])>0.5:
par_Rat10_count+=1
if np.max(par_iou[0:20])>0.5:
par_Rat20_count+=1
total_count+=1
print("---------------intact_bbox--------------------------")
print("IoU25_success_rate:",success_count_iou25/total_count)
print("IoU50_success_rate:", success_count_iou50 / total_count)
print("R@2:", Rat2_count / total_count)
print("R@5:", Rat5_count / total_count)
print("R@10:", Rat10_count / total_count)
print("R@20", Rat20_count / total_count)
print("Max_IoU",Max_IoU/total_count)
print("success mean IoU",iou_sum/success_count)
print("---------------partial_bbox--------------------------")
print("IoU25_success_rate:",par_success_count_iou25/total_count)
print("IoU50_success_rate:", par_success_count_iou50 / total_count)
print("R@2:", Rat2_count / total_count)
print("R@5:", par_Rat5_count / total_count)
print("R@10:", par_Rat10_count / total_count)
print("R@20", par_Rat20_count / total_count)
print("Max_IoU",par_Max_IoU/total_count)