-
-
Notifications
You must be signed in to change notification settings - Fork 2
/
evaluate_gpu.py
101 lines (88 loc) · 3.26 KB
/
evaluate_gpu.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
import scipy.io
import torch
import numpy as np
import time
#######################################################################
# Evaluate
def evaluate(qf,ql,qc,gf,gl,gc,junk_index1):
query = qf.view(-1,1)
# print(query.shape)
score = torch.mm(gf,query)
score = score.squeeze(1).cpu()
score = score.numpy()
# predict index
index = np.argsort(score) #from small to large
index = index[::-1]
# index = index[0:2000]
# good index
query_index = np.argwhere(gl==ql)
camera_index = np.argwhere(gc==qc)
good_index = np.setdiff1d(query_index, camera_index, assume_unique=True)
#junk_index1 = np.argwhere(gl==-1)
junk_index2 = np.intersect1d(query_index, camera_index)
junk_index = np.append(junk_index2, junk_index1) #.flatten())
CMC_tmp = compute_mAP(index, good_index, junk_index)
return CMC_tmp
def compute_mAP(index, good_index, junk_index):
ap = 0
cmc = torch.IntTensor(len(index)).zero_()
if good_index.size==0: # if empty
cmc[0] = -1
return ap,cmc
# remove junk_index
mask = np.in1d(index, junk_index, invert=True)
index = index[mask]
# find good_index index
ngood = len(good_index)
mask = np.in1d(index, good_index)
rows_good = np.argwhere(mask==True)
rows_good = rows_good.flatten()
cmc[rows_good[0]:] = 1
for i in range(ngood):
d_recall = 1.0/ngood
precision = (i+1)*1.0/(rows_good[i]+1)
if rows_good[i]!=0:
old_precision = i*1.0/rows_good[i]
else:
old_precision=1.0
ap = ap + d_recall*(old_precision + precision)/2
return ap, cmc
######################################################################
# main function
def main( query_path = './' ):
#result_q = scipy.io.loadmat(query_path+'/query_result_normal.mat')
result_q = scipy.io.loadmat(query_path+'/query_result.mat')
query_feature = torch.FloatTensor(result_q['query_f'])
query_cam = result_q['query_cam'][0]
query_label = result_q['query_label'][0]
result_g = scipy.io.loadmat('gallery_result.mat')
gallery_feature = torch.FloatTensor(result_g['gallery_f'])
gallery_cam = result_g['gallery_cam'][0]
gallery_label = result_g['gallery_label'][0]
query_feature = query_feature.cuda()
gallery_feature = gallery_feature.cuda()
CMC = torch.IntTensor(len(gallery_label)).zero_()
ap = 0.0
fail_index = []
junk_index1 = np.argwhere(gallery_label==-1) #not well-detected
#print(query_label)
for i in range(len(query_label)):
ap_tmp, CMC_tmp = evaluate(query_feature[i],query_label[i],query_cam[i],gallery_feature,gallery_label,gallery_cam, junk_index1)
if CMC_tmp[0]==-1:
continue
if CMC_tmp[0]==1: fail_index.append(i)
CMC += CMC_tmp
ap += ap_tmp
#print(i, CMC_tmp[0])
print(len(fail_index), fail_index)
CMC = CMC.float()
CMC = CMC/len(query_label) #average CMC
print('top1:%f top5:%f top10:%f mAP:%f'%(CMC[0],CMC[4],CMC[9],ap/len(query_label)))
save_result = (CMC[0],CMC[4],CMC[9],ap/len(query_label))
return save_result
if __name__=='__main__':
#since = time.time()
#query_path = './attack_query/baseline-9/16'
query_path = './'
main(query_path)
#print(time.time()-since)