def main():
cfg = Config()
TVT, TMO = set_devices(cfg.sys_device_ids)
#########
# Model #
#########
model = Model(net=cfg.net, pretrained=False)
#####################
# Load Model Weight #
#####################
used_file = cfg.model_weight_file or cfg.ckpt_file
loaded = torch.load(used_file, map_location=(lambda storage, loc: storage))
if cfg.model_weight_file == '':
loaded = loaded['state_dicts'][0]
load_state_dict(model, loaded)
print('Loaded model weights from {}'.format(used_file))
model = model.base
# Set eval mode. Force all BN layers to use global mean and variance, also disable dropout.
model.eval()
# Transfer Model to Specified Device.
TMO([model])
###################
# Extract Feature #
###################
im_dir = osp.expanduser(cfg.image_dir)
im_paths = get_im_names(im_dir,
pattern='*.jpg',
return_path=True,
return_np=False)
all_feat = []
import scipy.io as sci
for i, im_path in enumerate(im_paths):
im = np.asarray(Image.open(im_path).convert('RGB'))
im = pre_process_im(im, cfg)
im = Variable(TVT(torch.from_numpy(im).float()), volatile=True)
feat = model(im)
feat = feat.data.cpu().numpy()
all_feat.append(feat)
if (i + 1) % 100 == 0:
print('{}/{} images done'.format(i, len(im_paths)))
all_feat = np.concatenate(all_feat)
print('all_feat.shape:', all_feat.shape)
# all_feat = normalize(all_feat, axis=1)
# You can save your im_paths and features, or do anything else ...
sci.savemat(cfg.saved_feature_mat_file, {'features': all_feat})
def main():
cfg = Config()
TVT, TMO = set_devices(cfg.sys_device_ids)
#########
# Model #
#########
model = Model(net=cfg.net, pretrained=False)
#####################
# Load Model Weight #
#####################
used_file = cfg.model_weight_file or cfg.ckpt_file
loaded = torch.load(used_file, map_location=(lambda storage, loc: storage))
if cfg.model_weight_file == '':
loaded = loaded['state_dicts'][0]
load_state_dict(model, loaded)
print('Loaded model weights from {}'.format(used_file))
model = model.base
# Set eval mode. Force all BN layers to use global mean and variance, also disable dropout.
model.eval()
# Transfer Model to Specified Device.
TMO([model])
#######################
# Extract Image Crops #
#######################
all_feat = []
dataset = BoundingBoxCrops(cfg)
loader = DataLoader(dataset=dataset,
batch_size=cfg.batch_size,
shuffle=False,
num_workers=cfg.num_workers)
print('Processing detections')
t = tqdm(iter(loader), leave=False, total=len(loader))
for i, data in enumerate(t):
imgs, bad_boxes = data
imgs = Variable(TVT(imgs))
with torch.no_grad():
feats = model(imgs)
feats = feats.data.cpu().numpy()
for j in range(len(bad_boxes)):
if bad_boxes[j] == True:
feats[j,:] = np.zeros(1280)
sci.savemat(cfg.saved_feature_mat_path+'/'+"{:05d}".format(i)+'.mat', {'features':feats})
def main():
cfg = Config()
TVT, TMO = set_devices(cfg.sys_device_ids)
#########
# Model #
#########
model = Model(last_conv_stride=cfg.last_conv_stride)
# Set eval mode. Force all BN layers to use global mean and variance, also disable dropout.
model.eval()
# Transfer Model to Specified Device.
TMO([model])
#####################
# Load Model Weight #
#####################
used_file = cfg.model_weight_file or cfg.ckpt_file
loaded = torch.load(used_file, map_location=(lambda storage, loc: storage))
if cfg.model_weight_file == '':
loaded = loaded['state_dicts'][0]
load_state_dict(model, loaded)
print('Loaded model weights from {}'.format(used_file))
###################
# Extract Feature #
###################
im_dir = osp.expanduser('~/Dataset/market1501/Market-1501-v15.09.15/query')
im_paths = get_im_names(im_dir,
pattern='*.jpg',
return_path=True,
return_np=False)
all_feat = []
for i, im_path in enumerate(im_paths):
im = np.asarray(Image.open(im_path).convert('RGB'))
im = pre_process_im(im, cfg)
im = Variable(TVT(torch.from_numpy(im).float()), volatile=True)
feat = model(im)
feat = feat.data.cpu().numpy()
all_feat.append(feat)
if (i + 1) % 100 == 0:
print('{}/{} images done'.format(i, len(im_paths)))
all_feat = np.concatenate(all_feat)
print('all_feat.shape:', all_feat.shape)
all_feat = normalize(all_feat, axis=1)
def main():
cfg = Config()
TVT, TMO = set_devices(cfg.sys_device_ids)
#########
# Model #
#########
model = Model(last_conv_stride=cfg.last_conv_stride)
# Set eval mode. Force all BN layers to use global mean and variance, also disable dropout.
model.eval()
# Transfer Model to Specified Device.
TMO([model])
#####################
# Load Model Weight #
#####################
used_file = cfg.model_weight_file or cfg.ckpt_file
loaded = torch.load(used_file, map_location=(lambda storage, loc: storage))
if cfg.model_weight_file == '':
loaded = loaded['state_dicts'][0]
load_state_dict(model, loaded)
print('Loaded model weights from {}'.format(used_file))
###################
# Extract Feature #
###################
im_dir = osp.expanduser('../../5Labeled Data/labeledNew')
im_paths = get_im_names(im_dir,
pattern='*.png',
return_path=True,
return_np=False)
currentData = {}
imagesKnown = set()
with open('../../7Data Features/dataFeatures.json') as json_file:
currentData = json.load(json_file)
for image in currentData["images"]:
imagesKnown.add(image["imageName"])
all_feat = []
for i, im_path in enumerate(im_paths):
im = np.asarray(Image.open(im_path).convert('RGB'))
im = pre_process_im(im, cfg)
im = Variable(TVT(torch.from_numpy(im).float()), volatile=True)
feat = model(im)
feat = feat.data.cpu().numpy()
all_feat.append(feat)
if (i + 1) % 100 == 0:
print('{}/{} images done'.format(i, len(im_paths)))
all_feat = np.concatenate(all_feat)
print('all_feat.shape:', all_feat.shape)
all_feat = normalize(all_feat, axis=1)
# You can save your im_paths and features, or do anything else ...
for i, im_path in enumerate(im_paths):
if (im_path in imagesKnown):
print("image", im_path, "already exists")
pass
else:
currentData['images'].append({
'imageName': im_path,
'features': all_feat[i].tolist()
})
if (i + 1) % 100 == 0:
print('{}/{} images saved'.format(i, len(im_paths)))
with open('../../7Data Features/dataFeatures.json', 'w') as outfile:
json.dump(currentData, outfile)
def main():
cfg = Config()
# Redirect logs to both console and file.
if cfg.log_to_file:
ReDirectSTD(cfg.stdout_file, 'stdout', False)
ReDirectSTD(cfg.stderr_file, 'stderr', False)
TVT, TMO = set_devices(cfg.sys_device_ids)
# Dump the configurations to log.
import pprint
print('-' * 60)
print('cfg.__dict__')
pprint.pprint(cfg.__dict__)
print('-' * 60)
###########
# Dataset #
###########
test_set = create_dataset(**cfg.test_set_kwargs)
#########
# Model #
#########
model = Model(last_conv_stride=cfg.last_conv_stride)
# Model wrapper
model_w = DataParallel(model)
# May Transfer Model to Specified Device.
TMO([model])
#####################
# Load Model Weight #
#####################
# To first load weights to CPU
map_location = (lambda storage, loc: storage)
used_file = cfg.model_weight_file or cfg.ckpt_file
loaded = torch.load(used_file, map_location=map_location)
if cfg.model_weight_file == '':
loaded = loaded['state_dicts'][0]
load_state_dict(model, loaded)
print('Loaded model weights from {}'.format(used_file))
###################
# Extract Feature #
###################
test_set.set_feat_func(ExtractFeature(model_w, TVT))
with measure_time('Extracting feature...', verbose=True):
feat, ids, cams, im_names, marks = test_set.extract_feat(True, verbose=True)
#######################
# Select Query Images #
#######################
# Fix some query images, so that the visualization for different models can
# be compared.
# Sort in the order of image names
inds = np.argsort(im_names)
feat, ids, cams, im_names, marks = \
feat[inds], ids[inds], cams[inds], im_names[inds], marks[inds]
# query, gallery index mask
is_q = marks == 0
is_g = marks == 1
prng = np.random.RandomState(1)
# selected query indices
sel_q_inds = prng.permutation(range(np.sum(is_q)))[:cfg.num_queries]
q_ids = ids[is_q][sel_q_inds]
q_cams = cams[is_q][sel_q_inds]
q_feat = feat[is_q][sel_q_inds]
q_im_names = im_names[is_q][sel_q_inds]
####################
# Compute Distance #
####################
# query-gallery distance
q_g_dist = compute_dist(q_feat, feat[is_g], type='euclidean')
###########################
# Save Rank List as Image #
###########################
q_im_paths = [ospj(test_set.im_dir, n) for n in q_im_names]
save_paths = [ospj(cfg.exp_dir, 'rank_lists', n) for n in q_im_names]
g_im_paths = [ospj(test_set.im_dir, n) for n in im_names[is_g]]
for dist_vec, q_id, q_cam, q_im_path, save_path in zip(
q_g_dist, q_ids, q_cams, q_im_paths, save_paths):
rank_list, same_id = get_rank_list(
dist_vec, q_id, q_cam, ids[is_g], cams[is_g], cfg.rank_list_size)
save_rank_list_to_im(rank_list, same_id, q_im_path, g_im_paths, save_path)
def main():
cfg = Config()
# Redirect logs to both console and file.
if cfg.log_to_file:
ReDirectSTD(cfg.stdout_file, 'stdout', False)
ReDirectSTD(cfg.stderr_file, 'stderr', False)
TVT, TMO = set_devices(cfg.sys_device_ids)
# Dump the configurations to log.
import pprint
print('-' * 60)
print('cfg.__dict__')
pprint.pprint(cfg.__dict__)
print('-' * 60)
###########
# Dataset #
###########
test_set = create_dataset(**cfg.test_set_kwargs)
#########
# Model #
#########
model = Model(last_conv_stride=cfg.last_conv_stride)
# Model wrapper
model_w = DataParallel(model)
# May Transfer Model to Specified Device.
TMO([model])
#####################
# Load Model Weight #
#####################
# To first load weights to CPU
map_location = (lambda storage, loc: storage)
used_file = cfg.model_weight_file or cfg.ckpt_file
loaded = torch.load(used_file, map_location=map_location)
if cfg.model_weight_file == '':
loaded = loaded['state_dicts'][0]
load_state_dict(model, loaded)
print('Loaded model weights from {}'.format(used_file))
###################
# Extract Feature #
###################
test_set.set_feat_func(ExtractFeature(model_w, TVT))
with measure_time('Extracting feature...', verbose=True):
feat, ids, cams, im_names, marks = test_set.extract_feat(True, verbose=True)
#######################
# Select Query Images #
#######################
# Fix some query images, so that the visualization for different models can
# be compared.
# Sort in the order of image names
inds = np.argsort(im_names)
feat, ids, cams, im_names, marks = \
feat[inds], ids[inds], cams[inds], im_names[inds], marks[inds]
# query, gallery index mask
is_q = marks == 0
is_g = marks == 1
prng = np.random.RandomState(1)
# selected query indices
sel_q_inds = prng.permutation(range(np.sum(is_q)))[:cfg.num_queries]
q_ids = ids[is_q][sel_q_inds]
q_cams = cams[is_q][sel_q_inds]
q_feat = feat[is_q][sel_q_inds]
q_im_names = im_names[is_q][sel_q_inds]
####################
# Compute Distance #
####################
# query-gallery distance
f= open("../../4Scripts/Data Labeling Process/imagesDesc2.txt","w+")
#for controlIndex in range(len(im_names[is_q])):
for controlIndex in range(0,6):
f.write(im_names[is_q][controlIndex] + "|" + "|".join(list(map(str, feat[is_q][controlIndex])))+"\n")
#for controlIndex in range(len(im_names[is_g])):
for controlIndex in range(0,6):
f.write(im_names[is_g][controlIndex] + "|" + "|".join(list(map(str, feat[is_g][controlIndex])))+"\n")
f.close()