def get_reconstruction_objective_values(dp: DataProvider,
image_type: ImageType, t):
"""
This function dumps the output of the objective function given a permutation from the greedy comparator
:return: a tuple of 2 lists of log probabilities (scalars)
"""
solver = image_type_to_solver_with_comparator[image_type]
comparator = solver._t_to_comparator[t]
inputs = dp.get_fish_images(
) if image_type == ImageType.IMAGES else dp.get_docs_images()
inputs = [Shredder.shred(im, t, shuffle_shreds=False) for im in inputs]
correct_reconstruction_probas = []
incorrect_reconstruction_probas = []
for i, stacked_shreds in enumerate(inputs):
print('#{}-{}-{}'.format(i, image_type, t))
ltr_adj_probas, ttb_adj_probas = AdjacencyMatrixBuilder.build_adjacency_matrices(
comparator, stacked_shreds)
_, correct_log_objective = ObjectiveFunction.compute(
np.arange(t**2), ltr_adj_probas, ttb_adj_probas)
predicted_permutation, log_objective = solver.predict(
stacked_shreds, return_log_objective=True)
if np.array_equal(predicted_permutation, np.arange(t**2)):
correct_reconstruction_probas.append(log_objective)
else:
incorrect_reconstruction_probas.append(
(log_objective, correct_log_objective))
return correct_reconstruction_probas, incorrect_reconstruction_probas
def prepare_data(self, num_samples, resize=None):
num_samples = 2000
dp = DataProvider()
ts = (1, 2, 4, 5)
fish = dp.get_fish_images(num_samples=num_samples, resize=resize)
docs = dp.get_docs_images(num_samples=num_samples, resize=resize)
fish = list(itertools.chain(*[shred_shuffle_and_reconstruct(fish, t) for t in ts]))
docs = list(itertools.chain(*[shred_shuffle_and_reconstruct(docs, t) for t in ts]))
if resize is not None:
fish = list_of_images_to_numpy(fish)
docs = list_of_images_to_numpy(docs)
return fish, docs
def get_adjacent_crops_probabilities(dp: DataProvider, image_type: ImageType,
t):
"""
This function calculates the output of the comparator on adjacent crops
:return: list of probabilities (scalars)
"""
comparator = image_type_to_t_to_comparator[image_type][t]
inputs = dp.get_fish_images(
) if image_type == ImageType.IMAGES else dp.get_docs_images()
inputs = shred_and_resize_to(inputs, t,
(comparator.width, comparator.height))
adj_probabilities = []
for stacked_shreds in inputs:
for left_idx in range(t**2):
right_idx = left_idx + 1
if left_idx % t == t - 1:
continue
softmax = comparator.predict_is_left_probability(
[stacked_shreds[left_idx]], [stacked_shreds[right_idx]])
adj_probabilities.append(softmax[0][1])
return adj_probabilities
def get_non_adjacent_crops_probabilities(dp: DataProvider,
image_type: ImageType, t):
"""
This function calculates the output of the comparator on non adjacent crops
:return: list of probabilities (scalars)
"""
comparator = image_type_to_t_to_comparator[image_type][t]
inputs = dp.get_fish_images(
) if image_type == ImageType.IMAGES else dp.get_docs_images()
inputs = shred_and_resize_to(inputs, t,
(comparator.width, comparator.height))
non_adj_probabilities = []
for stacked_shreds in inputs:
left_idx, right_idx = 0, 1
while left_idx + 1 == right_idx:
left_idx, right_idx = tuple(
np.random.choice(t**2, 2, replace=False))
softmax = comparator.predict_is_left_probability(
[stacked_shreds[left_idx]], [stacked_shreds[right_idx]])
non_adj_probabilities.append(softmax[0][1])
return non_adj_probabilities
def get_number_of_images_with_same_patches_and_number_of_same_patches(
data_provider: DataProvider,
image_type: ImageType,
t: int,
width_height=None):
inputs = data_provider.get_fish_images() if image_type == ImageType.IMAGES else data_provider.get_docs_images()
if width_height is None:
inputs = list(map(lambda image: Shredder.shred(image, t), inputs))
else:
inputs = shred_and_resize_to(inputs, t, width_height)
number_of_pictures_with_same_patches = 0
number_of_patches_with_similar_in_same_picture = 0
for stacked_shreds in inputs:
picture_has_similar_patches = False
for left_shred in range(t**2):
picture_has_similar_to_this_shred = False
for right_shred in range(t**2):
if left_shred != right_shred and np.all(stacked_shreds[left_shred] == stacked_shreds[right_shred]):
picture_has_similar_to_this_shred = True
if picture_has_similar_to_this_shred:
picture_has_similar_patches = True
number_of_patches_with_similar_in_same_picture += 1
if picture_has_similar_patches:
number_of_pictures_with_same_patches += 1
return \
number_of_pictures_with_same_patches, \
number_of_patches_with_similar_in_same_picture, \
len(inputs), \
len(inputs) * (t ** 2)