def plshface(args, parallel_pool):
matrix_x = []
matrix_y = []
plotting_labels = []
plotting_scores = []
splits = []
print('>> EXPLORING DATASET')
dataset_dict = {value: index for index, value in enumerate(list_of_paths)}
dataset_list = zip(list_of_paths, list_of_labels)
dataset_list = set_maximum_samples(dataset_list, number_of_samples=SAMPLES)
# Split dataset into disjoint sets in terms of subjects and samples
known_tuples, unknown_tuples = split_known_unknown_sets(
dataset_list, known_set_size=KNOWN_SET_SIZE)
known_train, known_test = split_train_test_sets(
known_tuples, train_set_size=TRAIN_SET_SIZE)
to_discard, unknown_test = split_train_test_sets(
unknown_tuples, train_set_size=TRAIN_SET_SIZE)
print('>> LOADING GALLERY: {0} samples'.format(len(known_train)))
for counterA, gallery_sample in enumerate(known_train):
sample_path = gallery_sample[0]
sample_name = gallery_sample[1]
sample_index = dataset_dict[sample_path]
feature_vector = list_of_features[sample_index]
# create list of feature vectors and their corresponding target values
matrix_x.append(feature_vector)
matrix_y.append(sample_name)
# print(counterA, sample_path, sample_name)
print('>> SPLITTING POSITIVE/NEGATIVE SETS')
individuals = list(set(matrix_y))
os_cmc_score = np.zeros(len(individuals))
oaa_cmc_score = np.zeros(len(individuals))
for index in range(0, NUM_HASH):
splits.append(generate_pos_neg_dict(individuals))
# Converting list to numpy arrays
numpy_x = np.array(matrix_x)
numpy_y = np.array(matrix_y)
numpy_s = np.array(splits)
print('>> LEARNING OPEN-SET PLS MODELS:')
os_models = parallel_pool(
delayed(learn_plsh_model)(numpy_x, numpy_y, split)
for split in numpy_s)
print('>> LEARNING CLOSED-SET ONE-AGAINST-ALL PLS MODELS:')
oaa_splits = generate_oaa_splits(numpy_y)
oaa_models = parallel_pool(
delayed(learn_oaa_pls)(numpy_x, split) for split in oaa_splits)
print('>> LOADING KNOWN PROBE: {0} samples'.format(len(known_test)))
for counterB, probe_sample in enumerate(known_test):
# Obtaining probe feature vector and corresponding identity
sample_path = probe_sample[0]
sample_name = probe_sample[1]
sample_index = dataset_dict[sample_path]
feature_vector = list_of_features[sample_index]
# Projecting feature vector to every os_model
vote_dict = dict(map(lambda vote: (vote, 0), individuals))
for model in os_models:
pos_list = [
key for key, value in model[1].iteritems() if value == 1
]
response = model[0].predict_confidence(feature_vector)
for pos in pos_list:
vote_dict[pos] += response
# Sort open-set vote-list histogram
vote_list = vote_dict.items()
vote_list.sort(key=lambda tup: tup[1], reverse=True)
denominator = np.absolute(
np.mean([vote_list[1][1], vote_list[2][1], vote_list[3][1]]))
vote_ratio = vote_list[0][
1] / denominator if denominator > 0 else vote_list[0][1]
# Computer cmc score for open-set classification
for outer in range(0, len(individuals)):
for inner in range(0, outer + 1):
if vote_list[inner][0] == sample_name:
os_cmc_score[outer] += 1
break
# Sort closed-set responses
ooa_mls_responses = [
model[0].predict_confidence(feature_vector) for model in oaa_models
]
ooa_mls_labels = [model[1] for model in oaa_models]
responses = zip(ooa_mls_labels, ooa_mls_responses)
responses.sort(key=lambda tup: tup[1], reverse=True)
# Computer cmc score for closed-set classification
for outer in range(0, len(individuals)):
for inner in range(0, outer + 1):
if responses[inner][0] == sample_name:
oaa_cmc_score[outer] += 1
break
# Getting known set plotting relevant information
plotting_labels.append([(sample_name, 1.0)])
plotting_scores.append([(sample_name, vote_ratio)])
# print(counterB, sample_name, vote_ratio, vote_list[0][0], responses[0][0])
print('>> LOADING UNKNOWN PROBE: {0} samples'.format(len(unknown_tuples)))
for counterC, probe_sample in enumerate(unknown_test):
# Obtaining probe feature vector and corresponding identity
sample_path = probe_sample[0]
sample_name = probe_sample[1]
sample_index = dataset_dict[sample_path]
feature_vector = list_of_features[sample_index]
# Projecting feature vector to every os_model
vote_dict = dict(map(lambda vote: (vote, 0), individuals))
for model in os_models:
pos_list = [
key for key, value in model[1].iteritems() if value == 1
]
response = model[0].predict_confidence(feature_vector)
for pos in pos_list:
vote_dict[pos] += response
# Sort open-set vote-list histogram
vote_list = vote_dict.items()
vote_list.sort(key=lambda tup: tup[1], reverse=True)
denominator = np.absolute(
np.mean([vote_list[1][1], vote_list[2][1], vote_list[3][1]]))
vote_ratio = vote_list[0][
1] / denominator if denominator > 0 else vote_list[0][1]
# Getting unknown set plotting relevant information
plotting_labels.append([(sample_name, -1.0)])
plotting_scores.append([(sample_name, vote_ratio)])
# print(counterC, sample_name, vote_ratio, vote_list[0][0])
del os_models[:]
del oaa_models[:]
os_cmc = np.divide(os_cmc_score, len(known_test))
oaa_cmc = np.divide(oaa_cmc_score, len(known_test))
det = generate_det_curve(plotting_labels, plotting_scores)
pr = generate_precision_recall(plotting_labels, plotting_scores)
roc = generate_roc_curve(plotting_labels, plotting_scores)
fscore = compute_fscore(pr)
return os_cmc, oaa_cmc, det, pr, roc, fscore
def svmhface(args, parallel_pool):
PATH = str(args.path)
DATASET = str(args.file)
NUM_HASH = int(args.hash)
KNOWN_SET_SIZE = float(args.known_set_size)
TRAIN_SET_SIZE = float(args.train_set_size)
print('>> LOADING FEATURES FROM FILE')
with open(PATH + DATASET, 'rb') as input_file:
list_of_paths, list_of_labels, list_of_features = pickle.load(
input_file)
matrix_x = []
matrix_y = []
plotting_labels = []
plotting_scores = []
splits = []
print('>> EXPLORING DATASET')
dataset_dict = {value: index for index, value in enumerate(list_of_paths)}
dataset_list = zip(list_of_paths, list_of_labels)
known_tuples, unknown_tuples = split_known_unknown_sets(
dataset_list, known_set_size=KNOWN_SET_SIZE)
known_train, known_test = split_train_test_sets(
known_tuples, train_set_size=TRAIN_SET_SIZE)
print('>> LOADING GALLERY: {0} samples'.format(len(known_train)))
counterA = 0
for gallery_sample in known_train:
sample_path = gallery_sample[0]
sample_name = gallery_sample[1]
sample_index = dataset_dict[sample_path]
feature_vector = list_of_features[sample_index]
matrix_x.append(feature_vector)
matrix_y.append(sample_name)
counterA += 1
print(counterA, sample_path, sample_name)
print('>> SPLITTING POSITIVE/NEGATIVE SETS')
individuals = list(set(matrix_y))
cmc_score = np.zeros(len(individuals))
for index in range(0, NUM_HASH):
splits.append(generate_pos_neg_dict(individuals))
print('>> LEARNING SVM MODELS:')
numpy_x = np.array(matrix_x)
numpy_y = np.array(matrix_y)
numpy_s = np.array(splits)
models = parallel_pool(
delayed(learn_svmh_model)(numpy_x, numpy_y, split)
for split in numpy_s)
print('>> LOADING KNOWN PROBE: {0} samples'.format(len(known_test)))
counterB = 0
for probe_sample in known_test:
sample_path = probe_sample[0]
sample_name = probe_sample[1]
sample_index = dataset_dict[sample_path]
feature_vector = list_of_features[sample_index]
vote_dict = dict(map(lambda vote: (vote, 0), individuals))
for model in models:
pos_list = [
key for key, value in model[1].iteritems() if value == 1
]
response = model[0].predict([feature_vector])
# print(response)
for pos in pos_list:
vote_dict[pos] += response
result = vote_dict.items()
result.sort(key=lambda tup: tup[1], reverse=True)
for outer in range(len(individuals)):
for inner in range(outer + 1):
if result[inner][0] == sample_name:
cmc_score[outer] += 1
break
counterB += 1
denominator = np.absolute(np.mean([result[1][1], result[2][1]]))
if denominator > 0:
output = result[0][1] / denominator
else:
output = result[0][1]
print(counterB, sample_name, result[0][0], output[0])
# Getting known set plotting relevant information
plotting_labels.append([(sample_name, 1)])
plotting_scores.append([(sample_name, output[0])])
print('>> LOADING UNKNOWN PROBE: {0} samples'.format(len(unknown_tuples)))
counterC = 0
for probe_sample in unknown_tuples:
sample_path = probe_sample[0]
sample_name = probe_sample[1]
sample_index = dataset_dict[sample_path]
feature_vector = list_of_features[sample_index]
vote_dict = dict(map(lambda vote: (vote, 0), individuals))
for model in models:
pos_list = [
key for key, value in model[1].iteritems() if value == 1
]
response = model[0].predict([feature_vector])
for pos in pos_list:
vote_dict[pos] += response
result = vote_dict.items()
result.sort(key=lambda tup: tup[1], reverse=True)
counterC += 1
denominator = np.absolute(np.mean([result[1][1], result[2][1]]))
if denominator > 0:
output = result[0][1] / denominator
else:
output = result[0][1]
print(counterC, sample_name, result[0][0], output[0])
# Getting unknown set plotting relevant information
plotting_labels.append([(sample_name, -1)])
plotting_scores.append([(sample_name, output[0])])
# cmc_score_norm = np.divide(cmc_score, counterA)
# generate_cmc_curve(cmc_score_norm, DATASET + '_' + str(NUM_HASH) + '_' + DESCRIPTOR)
del models[:]
del list_of_paths[:]
del list_of_labels[:]
del list_of_features[:]
pr = generate_precision_recall(plotting_labels, plotting_scores)
roc = generate_roc_curve(plotting_labels, plotting_scores)
return pr, roc
def svm_oneclass(args):
PATH = str(args.path)
DATASET = str(args.file)
DESCRIPTOR = str(args.desc)
NUM_HASH = int(args.hash)
IMG_WIDTH = int(args.width)
IMG_HEIGHT = int(args.height)
matrix_x = []
matrix_y = []
models = []
splits = []
nmatrix_x = []
nmatrix_y = []
x_train = []
y_train = []
nx_train = []
ny_train = []
plotting_labels = []
plotting_scores = []
vgg_model = None
if DESCRIPTOR == 'df':
from vggface import VGGFace
vgg_model = VGGFace()
print('>> EXPLORING DATASET')
dataset_list = load_txt_file(PATH + DATASET)
known_tuples, unknown_tuples = split_known_unknown_sets(dataset_list,
known_set_size=0.5)
known_train, known_test = split_train_test_sets(known_tuples,
train_set_size=0.5)
print(known_train)
counterA = 0
for gallery_sample in known_train:
sample_path = gallery_sample[0]
sample_name = gallery_sample[1]
gallery_path = PATH + sample_path
gallery_image = cv.imread(gallery_path, cv.IMREAD_COLOR)
if DESCRIPTOR == 'hog':
gallery_image = cv.resize(gallery_image, (IMG_HEIGHT, IMG_WIDTH))
feature_vector = Descriptor.get_hog(gallery_image)
elif DESCRIPTOR == 'df':
feature_vector = Descriptor.get_deep_feature(gallery_image,
vgg_model,
layer_name='fc6')
matrix_x.append(feature_vector)
matrix_y.append(sample_name)
counterA += 1
print(counterA, sample_path, sample_name)
print('>> GENERATING FILES TO SVM')
counterSVM = 0
for feature in matrix_x:
y_train.insert(counterSVM, 1)
x_train.insert(counterSVM, {})
count_inner = 0
for pos in feature:
x_train[counterSVM].update({count_inner: pos})
count_inner += 1
counterSVM += 1
print('>> GENERATING THE SVM MODEL')
x_train_total = x_train + nx_train
y_train_total = y_train + ny_train
besthit = 0
bestn = 0
bestg = 0
for n in range(1, 50):
for g in range(-15, 3):
nu = n / 100
gamma = pow(2, g)
parameters = '-s 2 -t 2'
parameters = parameters + ' -g ' + str(gamma) + ' -n ' + str(nu)
m = svm_train(y_train_total, x_train_total, parameters)
hits = 0
#print('>> LOADING KNOWN PROBE: {0} samples'.format(len(known_test)))
counterB = 0
for probe_sample in known_test:
sample_path = probe_sample[0]
sample_name = probe_sample[1]
query_path = PATH + sample_path
query_image = cv.imread(query_path, cv.IMREAD_COLOR)
if DESCRIPTOR == 'hog':
query_image = cv.resize(query_image,
(IMG_HEIGHT, IMG_WIDTH))
feature_vector = Descriptor.get_hog(query_image)
elif DESCRIPTOR == 'df':
feature_vector = Descriptor.get_deep_feature(
query_image, vgg_model)
count_inner = 0
x_teste = []
y_teste = []
y_teste.insert(0, 1)
x_teste.insert(0, {})
for pos in feature_vector:
x_teste[0].update({count_inner: pos})
count_inner += 1
p_label, p_acc, p_val = svm_predict(y_teste, x_teste, m)
counterB += 1
# Getting known set plotting relevant information
plotting_labels.append([(sample_name, 1)])
plotting_scores.append([(sample_name, p_label[0])])
if p_label[0] == 1:
hits = hits + 1
print('>> LOADING UNKNOWN PROBE: {0} samples'.format(
len(unknown_tuples)))
counterC = 0
for probe_sample in unknown_tuples:
sample_path = probe_sample[0]
sample_name = probe_sample[1]
query_path = PATH + sample_path
query_image = cv.imread(query_path, cv.IMREAD_COLOR)
if DESCRIPTOR == 'hog':
query_image = cv.resize(query_image,
(IMG_HEIGHT, IMG_WIDTH))
feature_vector = Descriptor.get_hog(query_image)
elif DESCRIPTOR == 'df':
feature_vector = Descriptor.get_deep_feature(
query_image, vgg_model)
count_inner = 0
x_teste = []
y_teste = []
y_teste.insert(0, -1)
x_teste.insert(0, {})
for pos in feature_vector:
x_teste[0].update({count_inner: pos})
count_inner += 1
p_label, p_acc, p_val = svm_predict(y_teste, x_teste, m)
counterC += 1
# Getting unknown set plotting relevant information
plotting_labels.append([(sample_name, -1)])
plotting_scores.append([(sample_name, p_label[0])])
if p_label[0] == -1:
hits = hits + 1
if hits > besthit:
besthit = hits
bestn = nu
bestg = gamma
# cmc_score_norm = np.divide(cmc_score, counterA)
# generate_cmc_curve(cmc_score_norm, DATASET + '_' + str(NUM_HASH) + '_' + DESCRIPTOR)
print(besthits)
print(bestn)
print(bestg)
pr = generate_precision_recall(plotting_labels, plotting_scores)
roc = generate_roc_curve(plotting_labels, plotting_scores)
return pr, roc
def fcnhface(args, parallel_pool):
matrix_x = []
matrix_y = []
plotting_labels = []
plotting_scores = []
models = []
splits = []
print('>> EXPLORING DATASET')
dataset_dict = {value: index for index, value in enumerate(list_of_paths)}
dataset_list = zip(list_of_paths, list_of_labels)
dataset_list = set_maximum_samples(dataset_list, number_of_samples=SAMPLES)
known_tuples, unknown_tuples = split_known_unknown_sets(
dataset_list, known_set_size=KNOWN_SET_SIZE)
known_train, known_test = split_train_test_sets(
known_tuples, train_set_size=TRAIN_SET_SIZE)
print('>> LOADING GALLERY: {0} samples'.format(len(known_train)))
counterA = 0
for gallery_sample in known_train:
sample_path = gallery_sample[0]
sample_name = gallery_sample[1]
sample_index = dataset_dict[sample_path]
feature_vector = list_of_features[sample_index]
matrix_x.append(feature_vector)
matrix_y.append(sample_name)
counterA += 1
# print(counterA, sample_path, sample_name)
print('>> SPLITTING POSITIVE/NEGATIVE SETS')
individuals = list(set(matrix_y))
cmc_score = np.zeros(len(individuals))
for index in range(0, NUM_HASH):
splits.append(generate_pos_neg_dict(individuals))
print('>> LEARNING FC MODELS:')
numpy_x = np.array(matrix_x)
numpy_y = np.array(matrix_y)
numpy_s = np.array(splits)
# models = [learn_fc_model(numpy_x, numpy_y, split) for split in numpy_s]
models = parallel_pool(
delayed(learn_fc_model)(numpy_x, numpy_y, split) for split in numpy_s)
print('>> LOADING KNOWN PROBE: {0} samples'.format(len(known_test)))
counterB = 0
for probe_sample in known_test:
sample_path = probe_sample[0]
sample_name = probe_sample[1]
sample_index = dataset_dict[sample_path]
feature_vector = np.array(list_of_features[sample_index])
vote_dict = dict(map(lambda vote: (vote, 0), individuals))
for k in range(0, len(models)):
pos_list = [
key for key, value in models[k][1].iteritems() if value == 1
]
pred = models[k][0].predict(
feature_vector.reshape(1, feature_vector.shape[0]))
response = pred[0][1]
#print(response)
for pos in pos_list:
vote_dict[pos] += response
result = vote_dict.items()
result.sort(key=lambda tup: tup[1], reverse=True)
for outer in range(len(individuals)):
for inner in range(outer + 1):
if result[inner][0] == sample_name:
cmc_score[outer] += 1
break
counterB += 1
denominator = np.absolute(np.mean([result[1][1], result[2][1]]))
if denominator > 0:
output = result[0][1] / denominator
else:
output = result[0][1]
# print(counterB, sample_name, result[0][0], output)
# Getting known set plotting relevant information
plotting_labels.append([(sample_name, 1)])
plotting_scores.append([(sample_name, output)])
print('>> LOADING UNKNOWN PROBE: {0} samples'.format(len(unknown_tuples)))
counterC = 0
for probe_sample in unknown_tuples:
sample_path = probe_sample[0]
sample_name = probe_sample[1]
sample_index = dataset_dict[sample_path]
feature_vector = np.array(list_of_features[sample_index])
vote_dict = dict(map(lambda vote: (vote, 0), individuals))
#print (vote_dict)
for k in range(0, len(models)):
pos_list = [
key for key, value in models[k][1].iteritems() if value == 1
]
pred = models[k][0].predict(
feature_vector.reshape(1, feature_vector.shape[0]))
response = pred[0][1]
for pos in pos_list:
vote_dict[pos] += response
result = vote_dict.items()
result.sort(key=lambda tup: tup[1], reverse=True)
counterC += 1
denominator = np.absolute(np.mean([result[1][1], result[2][1]]))
if denominator > 0:
output = result[0][1] / denominator
else:
output = result[0][1]
# print(counterC, sample_name, result[0][0], output)
# Getting unknown set plotting relevant information
plotting_labels.append([(sample_name, -1)])
plotting_scores.append([(sample_name, output)])
# cmc_score_norm = np.divide(cmc_score, counterA)
# generate_cmc_curve(cmc_score_norm, DATASET + '_' + str(NUM_HASH) + '_' + DESCRIPTOR)
del models[:]
pr = generate_precision_recall(plotting_labels, plotting_scores)
roc = generate_roc_curve(plotting_labels, plotting_scores)
fscore = compute_fscore(pr)
return pr, roc, fscore
def plshface(args):
PATH = str(args.path)
DATASET = str(args.file)
DESCRIPTOR = str(args.desc)
NUM_HASH = int(args.hash)
IMG_WIDTH = int(args.width)
IMG_HEIGHT = int(args.height)
TRAIN_SET_SIZE = float(args.train_set_size)
matrix_x = []
matrix_y = []
splits = []
plotting_labels = []
plotting_scores = []
vgg_model = None
if DESCRIPTOR == 'df':
vgg_model = VGGFace()
print('>> EXPLORING DATASET')
dataset_list = load_txt_file(PATH + DATASET)
known_train, known_test = split_train_test_sets(
dataset_list, train_set_size=TRAIN_SET_SIZE)
print('>> LOADING GALLERY: {0} samples'.format(len(known_train)))
counterA = 0
for gallery_sample in known_train:
sample_path = gallery_sample[0]
sample_name = gallery_sample[1]
gallery_path = PATH + sample_path
gallery_image = cv.imread(gallery_path, cv.IMREAD_COLOR)
if DESCRIPTOR == 'hog':
gallery_image = cv.resize(gallery_image, (IMG_HEIGHT, IMG_WIDTH))
feature_vector = Descriptor.get_hog(gallery_image)
elif DESCRIPTOR == 'df':
feature_vector = Descriptor.get_deep_feature(gallery_image,
vgg_model,
layer_name='fc6')
matrix_x.append(feature_vector)
matrix_y.append(sample_name)
counterA += 1
print(counterA, sample_path, sample_name)
print('>> SPLITTING POSITIVE/NEGATIVE SETS')
individuals = list(set(matrix_y))
cmc_score = np.zeros(len(individuals))
for index in range(0, NUM_HASH):
splits.append(generate_pos_neg_dict(individuals))
print('>> LEARNING PLS MODELS:')
input_list = itertools.izip(splits, itertools.repeat((matrix_x, matrix_y)))
models = Parallel(n_jobs=1, verbose=11,
backend='threading')(map(delayed(learn_plsh_model),
input_list))
print('>> LOADING KNOWN PROBE: {0} samples'.format(len(known_test)))
counterB = 0
for probe_sample in known_test:
sample_path = probe_sample[0]
sample_name = probe_sample[1]
query_path = PATH + sample_path
query_image = cv.imread(query_path, cv.IMREAD_COLOR)
if DESCRIPTOR == 'hog':
query_image = cv.resize(query_image, (IMG_HEIGHT, IMG_WIDTH))
feature_vector = Descriptor.get_hog(query_image)
elif DESCRIPTOR == 'df':
feature_vector = Descriptor.get_deep_feature(
query_image, vgg_model)
vote_dict = dict(map(lambda vote: (vote, 0), individuals))
for model in models:
pos_list = [
key for key, value in model[1].iteritems() if value == 1
]
response = model[0].predict_confidence(feature_vector)
for pos in pos_list:
vote_dict[pos] += response
result = vote_dict.items()
result.sort(key=lambda tup: tup[1], reverse=True)
for outer in range(len(individuals)):
for inner in range(outer + 1):
if result[inner][0] == sample_name:
cmc_score[outer] += 1
break
counterB += 1
denominator = np.absolute(np.mean([result[1][1], result[2][1]]))
if denominator > 0:
output = result[0][1] / denominator
else:
output = result[0][1]
print(counterB, sample_name, result[0][0], output)
# Getting known set plotting relevant information
plotting_labels.append([(sample_name, 1)])
plotting_scores.append([(sample_name, output)])
cmc_score_norm = np.divide(cmc_score, counterA)
return cmc_score_norm