def evaluate_all(distmat,
query=None,
gallery=None,
query_ids=None,
gallery_ids=None,
query_cams=None,
gallery_cams=None,
cmc_topk=(1, 5, 10)):
if query is not None and gallery is not None:
query_ids = [pid for _, pid, _ in query]
gallery_ids = [pid for _, pid, _ in gallery]
query_cams = [cam for _, _, cam in query]
gallery_cams = [cam for _, _, cam in gallery]
else:
assert (query_ids is not None and gallery_ids is not None
and query_cams is not None and gallery_cams is not None)
# Compute mean AP
mAP = mean_ap(distmat, query_ids, gallery_ids, query_cams, gallery_cams)
print('Mean AP: {:4.1%}'.format(mAP))
# Compute all kinds of CMC scores
cmc_configs = {
# 'allshots': dict(separate_camera_set=False,
# single_gallery_shot=False,
# first_match_break=False),
# 'cuhk03': dict(separate_camera_set=True,
# single_gallery_shot=True,
# first_match_break=False),
'market1501':
dict(separate_camera_set=False,
single_gallery_shot=False,
first_match_break=True)
}
cmc_scores = {
name: cmc(distmat, query_ids, gallery_ids, query_cams, gallery_cams,
**params)
for name, params in cmc_configs.items()
}
# print('CMC Scores{:>12}{:>12}{:>12}'
# .format('allshots', 'cuhk03', 'market1501'))
# for k in cmc_topk:
# print(' top-{:<4}{:12.1%}{:12.1%}{:12.1%}'
# .format(k, cmc_scores['allshots'][k - 1],
# cmc_scores['cuhk03'][k - 1],
# cmc_scores['market1501'][k - 1]))
# Use the allshots cmc top-1 score for validation criterion
return [cmc_scores['market1501'][k] for k in [0, 4, 9, 19]]
def evaluate_video(distmat,
query=None,
gallery=None,
query_ids=None,
gallery_ids=None,
query_cams=None,
gallery_cams=None,
cmc_topk=(1, 5, 10, 20),
dataset=None):
if query is not None and gallery is not None:
query_ids = [pid for _, pid, _ in query]
gallery_ids = [pid for _, pid, _ in gallery]
query_cams = [cam for _, _, cam in query]
gallery_cams = [cam for _, _, cam in gallery]
else:
assert (query_ids is not None and gallery_ids is not None
and query_cams is not None and gallery_cams is not None)
mAP = mean_ap(distmat, query_ids, gallery_ids, query_cams, gallery_cams)
print('Mean AP: {:4.1%}'.format(mAP))
# Compute all kinds of CMC scores
if True:
cmc_configs = {
'market1501':
dict(separate_camera_set=False,
single_gallery_shot=False,
first_match_break=True)
}
cmc_scores = {
name: cmc(distmat, query_ids, gallery_ids, query_cams,
gallery_cams, **params)
for name, params in cmc_configs.items()
}
print('CMC Scores{:>12}'.format('market1501'))
for k in cmc_topk:
print(' top-{:<4}{:12.1%}'.format(k, cmc_scores['market1501'][k -
1]))
return cmc_scores['market1501'][0], mAP
def evaluate_1st(self, distmat, query=None, gallery=None):
if query is not None and gallery is not None:
query_ids = [pid for _, pid, _ in query]
gallery_ids = [pid for _, pid, _ in gallery]
query_cams = [cam for _, _, cam in query]
gallery_cams = [cam for _, _, cam in gallery]
else:
raise RuntimeError(
'please provide the query and gallery information')
distmat = to_numpy(distmat)
mAP = mean_ap(distmat, query_ids, gallery_ids, query_cams,
gallery_cams)
print('Mean AP: {:4.1%}'.format(mAP))
indices = np.argsort(distmat, axis=1)
indices = np.argsort(indices, axis=1)
mask = (indices < self.select_size).astype(float)
cmc_configs = {
# 'allshots': dict(separate_camera_set=False,
# single_gallery_shot=False,
# first_match_break=False),
# 'cuhk03': dict(separate_camera_set=True,
# single_gallery_shot=True,
# first_match_break=False),
'market1501':
dict(separate_camera_set=False,
single_gallery_shot=False,
first_match_break=True)
}
cmc_scores = {
name: cmc(distmat, query_ids, gallery_ids, query_cams,
gallery_cams, **params)
for name, params in cmc_configs.items()
}
return [cmc_scores['market1501'][k] for k in [0, 4, 9, 19]], mask
def evaluate_all(distmat,
query=None,
gallery=None,
query_ids=None,
gallery_ids=None,
query_cams=None,
gallery_cams=None,
cmc_topk=(1, 5, 10)):
if query is not None and gallery is not None:
query_ids = [pid for _, pid, _ in query]
gallery_ids = [pid for _, pid, _ in gallery]
query_cams = [cam for _, _, cam in query]
gallery_cams = [cam for _, _, cam in gallery]
else:
assert (query_ids is not None and gallery_ids is not None
and query_cams is not None and gallery_cams is not None)
# Compute mean AP
print(distmat)
distnp = distmat.cpu().numpy()
np.save('dist_demo', distnp)
mAP = mean_ap(distmat, query_ids, gallery_ids, query_cams, gallery_cams)
print('Mean AP: {:4.1%}'.format(mAP))
# Compute all kinds of CMC scores
cmc_configs = {
# 'allshots': dict(separate_camera_set=False,
# single_gallery_shot=False,
# first_match_break=False),
# 'cuhk03': dict(separate_camera_set=True,
# single_gallery_shot=True,
# first_match_break=False),
'market1501':
dict(separate_camera_set=False,
single_gallery_shot=False,
first_match_break=True)
}
cmc_scores = {
name: cmc(distmat, query_ids, gallery_ids, query_cams, gallery_cams,
**params)
for name, params in cmc_configs.items()
}
# print('CMC Scores{:>12}{:>12}{:>12}'
# .format('allshots', 'cuhk03', 'market1501'))
# for k in cmc_topk:
# print(' top-{:<4}{:12.1%}{:12.1%}{:12.1%}'
# .format(k, cmc_scores['allshots'][k - 1],
# cmc_scores['cuhk03'][k - 1],
# cmc_scores['market1501'][k - 1]))
# Use the allshots cmc top-1 score for validation criterion
print([cmc_scores['market1501'][k] for k in [0, 4, 9, 19]])
############### multiple query-images #######################
joint_query = list(zip(query_ids, query_cams))
summarize_index = list(set(joint_query))
query_num = len(summarize_index)
gallery_num = len(gallery_ids)
sum_distmat = torch.zeros(query_num, gallery_num).cuda()
for newind, sumind in enumerate(summarize_index):
sum_indpos = [
posx for posx, x in enumerate(joint_query) if x == sumind
]
### mean results
select_results = torch.index_select(
distmat, 0,
torch.LongTensor(sum_indpos).cuda())
select_mean = torch.mean(select_results, 0)
sum_distmat[newind] = select_mean
## QUERY CAMERA
multiquery_ids, multiquery_cams = zip(*summarize_index)
multiquery_ids = list(multiquery_ids)
multiquery_cams = list(multiquery_cams)
multimAP = mean_ap(sum_distmat, multiquery_ids, gallery_ids,
multiquery_cams, gallery_cams)
print('Mean AP: {:4.1%}'.format(multimAP))
# Compute all kinds of CMC scores
multicmc_configs = {
# 'allshots': dict(separate_camera_set=False,
# single_gallery_shot=False,
# first_match_break=False),
# 'cuhk03': dict(separate_camera_set=True,
# single_gallery_shot=True,
# first_match_break=False),
'market1501':
dict(separate_camera_set=False,
single_gallery_shot=False,
first_match_break=True)
}
multicmc_scores = {
name: cmc(sum_distmat, multiquery_ids, gallery_ids, multiquery_cams,
gallery_cams, **params)
for name, params in multicmc_configs.items()
}
# print('CMC Scores{:>12}{:>12}{:>12}'
# .format('allshots', 'cuhk03', 'market1501'))
# for k in cmc_topk:
# print(' top-{:<4}{:12.1%}{:12.1%}{:12.1%}'
# .format(k, cmc_scores['allshots'][k - 1],
# cmc_scores['cuhk03'][k - 1],
# cmc_scores['market1501'][k - 1]))
# Use the allshots cmc top-1 score for validation criterion
print([multicmc_scores['market1501'][k] for k in [0, 4, 9, 19]])
return [cmc_scores['market1501'][k] for k in [0, 4, 9, 19]]
