def __init__(self, feature_extractor, gpu, batch_size):
self.feature_extractor = feature_extractor
self.gpu = gpu
self.batch_size = batch_size
global sess
sess = init_gpu(gpu)
global graph
graph = tf.get_default_graph()
model_dir = os.path.join(os.path.expanduser('~'), '.fawkes')
if not os.path.exists(os.path.join(model_dir, "mtcnn.p.gz")):
os.makedirs(model_dir, exist_ok=True)
get_file("mtcnn.p.gz",
"http://mirror.cs.uchicago.edu/fawkes/files/mtcnn.p.gz",
cache_dir=model_dir,
cache_subdir='')
self.fs_names = [feature_extractor]
if isinstance(feature_extractor, list):
self.fs_names = feature_extractor
self.aligner = aligner(sess)
self.feature_extractors_ls = [
load_extractor(name) for name in self.fs_names
]
self.protector = None
self.protector_param = None
def main():
sess = init_gpu(args.gpu)
ali = aligner(sess)
print("Build attacker's model")
image_paths = glob.glob(os.path.join(args.directory, "*"))
cloak_file_name = "low_cloaked"
original_image_paths = sorted(
[path for path in image_paths if "cloaked" not in path.split("/")[-1]])
protect_image_paths = sorted([
path for path in image_paths if cloak_file_name in path.split("/")[-1]
])
original_faces = Faces(original_image_paths,
ali,
verbose=1,
eval_local=True)
original_faces = original_faces.cropped_faces
cloaked_faces = Faces(protect_image_paths, ali, verbose=1, eval_local=True)
cloaked_faces = cloaked_faces.cropped_faces
if len(original_faces) <= 10:
raise Exception(
"Must have more than 10 protected images to run the evaluation")
num_classes = args.num_other_classes + 1
datagen = DataGenerator(original_faces, cloaked_faces)
original_test_X, original_test_Y = datagen.test_original()
print("{} Training Images | {} Testing Images".format(
len(datagen.protect_images_train), len(original_test_X)))
train_generator = datagen.generate()
test_generator = datagen.generate(test=True)
base_model = load_extractor(args.base_model)
model = load_victim_model(teacher_model=base_model,
number_classes=num_classes)
cb = CallbackGenerator(original_imgs=original_test_X,
protect_imgs=cloaked_faces,
original_y=original_test_Y,
original_protect_y=None,
test_gen=test_generator)
model.fit_generator(train_generator,
steps_per_epoch=num_classes * 10 // 32,
epochs=args.n_epochs,
verbose=2,
callbacks=[cb])
_, acc_original = model.evaluate(original_test_X,
original_test_Y,
verbose=0)
print("Protection Success Rate: {:.4f}".format(1 - acc_original))
def __init__(self, feature_extractor, gpu, batch_size):
global graph
graph = tf.get_default_graph()
self.feature_extractor = feature_extractor
self.gpu = gpu
self.batch_size = batch_size
self.sess = init_gpu(gpu)
self.aligner = aligner(self.sess)
self.fs_names = [feature_extractor]
if isinstance(feature_extractor, list):
self.fs_names = feature_extractor
self.feature_extractors_ls = [load_extractor(name) for name in self.fs_names]
def __init__(self, feature_extractor, gpu, batch_size, mode="low"):
self.feature_extractor = feature_extractor
self.gpu = gpu
self.batch_size = batch_size
self.mode = mode
th, max_step, lr, extractors = self.mode2param(self.mode)
self.th = th
self.lr = lr
self.max_step = max_step
if gpu is not None:
init_gpu(gpu)
self.aligner = aligner()
self.protector = None
self.protector_param = None
self.feature_extractors_ls = [load_extractor(name) for name in extractors]
def __init__(self, feature_extractor, gpu, batch_size, mode="low"):
self.feature_extractor = feature_extractor
self.gpu = gpu
self.batch_size = batch_size
self.mode = mode
th, max_step, lr, extractors = self.mode2param(self.mode)
self.th = th
self.lr = lr
self.max_step = max_step
init_gpu(gpu)
# model_dir = os.path.join(os.path.expanduser('~'), '.fawkes')
# if not os.path.exists(os.path.join(model_dir)):
# os.makedirs(model_dir, exist_ok=True)
self.aligner = aligner()
self.protector = None
self.protector_param = None
self.feature_extractors_ls = [
load_extractor(name) for name in extractors
]