def prediction_loss(self, truths, palette):
def spatial_loss(truth_features, predicted_features, space_desc):
feature_losses = []
for truth, prediction, spec in zip(truth_features,
predicted_features,
space_desc.features):
if spec.type == FeatureType.CATEGORICAL:
truth = tf.transpose(truth, (0, 2, 3, 1))
prediction = tf.transpose(prediction, (0, 2, 3, 1))
feature_losses.append(
tf.losses.softmax_cross_entropy(truth, prediction))
summary_image = tf.argmax(
tf.concat([truth, prediction], 2), 3)
summary_image = tf.gather(
palette[space_desc.index][spec.index], summary_image)
tf.summary.image(spec.name, summary_image)
else:
feature_losses.append(
tf.losses.mean_squared_error(truth, prediction))
summary_image = tf.concat([truth, prediction], 3)
tf.summary.image(spec.name,
tf.transpose(summary_image, (0, 2, 3, 1)))
tf.summary.scalar(spec.name, feature_losses[-1])
return tf.reduce_mean(tf.stack(feature_losses))
with tf.name_scope('prediction_loss'):
spatial_losses = []
for s in self.env_spec.spaces:
with tf.name_scope(s.name):
loss = spatial_loss(truths[s.index],
self.out_pred[s.index], s)
spatial_losses.append(loss)
tf.summary.scalar('loss', loss)
loss = tf.reduce_mean(tf.stack(spatial_losses))
tf.summary.scalar('loss', loss)
return loss
def compute_target_greedy_q(reward, gamma, next_actions, next_q_values,
next_states, terminals):
"""Computes the optimal target Q value with the adaptive greedy algorithm.
This algorithm corresponds to the method "GT" in
Ie et al. https://arxiv.org/abs/1905.12767..
Args:
reward: [batch_size] tensor, the immediate reward.
gamma: float, discount factor with the usual RL meaning.
next_actions: [batch_size, slate_size] tensor, the next slate.
next_q_values: [batch_size, num_of_documents] tensor, the q values of the
documents in the next step.
next_states: [batch_size, 1 + num_of_documents] tensor, the features for the
user and the docuemnts in the next step.
terminals: [batch_size] tensor, indicating if this is a terminal step.
Returns:
[batch_size] tensor, the target q values.
"""
slate_size = next_actions.get_shape().as_list()[1]
stack_number = -1
user_obs = next_states[:, 0, :, stack_number]
doc_obs = next_states[:, 1:, :, stack_number]
batch_size = next_q_values.get_shape().as_list()[0]
next_greedy_q_list = []
for i in range(batch_size):
s, s_no_click = score_documents_tf(user_obs[i], doc_obs[i])
q = next_q_values[i]
slate = select_slate_greedy(slate_size, s_no_click, s, q)
p_selected = compute_probs_tf(slate, s, s_no_click)
q_selected = tf.gather(q, slate)
next_greedy_q_list.append(
tf.reduce_sum(input_tensor=p_selected * q_selected))
next_greedy_q_values = tf.stack(next_greedy_q_list)
return reward + gamma * next_greedy_q_values * (
1. - tf.cast(terminals, tf.float32))
def select_slate_optimal(slate_size, s_no_click, s, q):
"""Selects the slate using exhaustive search.
This algorithm corresponds to the method "OS" in
Ie et al. https://arxiv.org/abs/1905.12767.
Args:
slate_size: int, the size of the recommendation slate.
s_no_click: float tensor, the score for not clicking any document.
s: [num_of_documents] tensor, the scores for clicking documents.
q: [num_of_documents] tensor, the predicted q values for documents.
Returns:
[slate_size] tensor, the selected slate.
"""
num_candidates = s.shape.as_list()[0]
# Obtain all possible slates given current docs in the candidate set.
mesh_args = [list(range(num_candidates))] * slate_size
slates = tf.stack(tf.meshgrid(*mesh_args), axis=-1)
slates = tf.reshape(slates, shape=(-1, slate_size))
# Filter slates that include duplicates to ensure each document is picked
# at most once.
unique_mask = tf.map_fn(
lambda x: tf.equal(tf.size(input=x), tf.size(input=tf.unique(x)[0])),
slates,
dtype=tf.bool)
slates = tf.boolean_mask(tensor=slates, mask=unique_mask)
slate_q_values = tf.gather(s * q, slates)
slate_scores = tf.gather(s, slates)
slate_normalizer = tf.reduce_sum(input_tensor=slate_scores,
axis=1) + s_no_click
slate_q_values = slate_q_values / tf.expand_dims(slate_normalizer, 1)
slate_sum_q_values = tf.reduce_sum(input_tensor=slate_q_values, axis=1)
max_q_slate_index = tf.argmax(input=slate_sum_q_values)
return tf.gather(slates, max_q_slate_index, axis=0)
def rotate_points_around_axis(points, rotation_angle, axis=2):
"""Rotates points around axis.
Args:
points: A tf.float32 tensor of size [N, 3] containing points.
rotation_angle: A float value containing the rotation angle in radians.
axis: A value in [0, 1, 2] for rotating around x, y, z axis.
Returns:
rotated_points: A tf.float32 tensor of size [N, 3] containing points.
"""
if axis not in [0, 1, 2]:
raise ValueError(('axis is out of bound: %d' % axis))
c = tf.cos(rotation_angle)
s = tf.sin(rotation_angle)
new_points = [points[:, 0], points[:, 1], points[:, 2]]
other_axis = list(set([0, 1, 2]) - set([axis]))
new_points[other_axis[0]] = (points[:, other_axis[0]] * c -
points[:, other_axis[1]] * s)
new_points[other_axis[1]] = (points[:, other_axis[0]] * s +
points[:, other_axis[1]] * c)
return tf.stack(new_points, axis=1)
def compute_target_sarsa(reward, gamma, next_actions, next_q_values,
next_states, terminals):
"""Computes the SARSA target Q value.
Args:
reward: [batch_size] tensor, the immediate reward.
gamma: float, discount factor with the usual RL meaning.
next_actions: [batch_size, slate_size] tensor, the next slate.
next_q_values: [batch_size, num_of_documents] tensor, the q values of the
documents in the next step.
next_states: [batch_size, 1 + num_of_documents] tensor, the features for the
user and the docuemnts in the next step.
terminals: [batch_size] tensor, indicating if this is a terminal step.
Returns:
[batch_size] tensor, the target q values.
"""
stack_number = -1
user_obs = next_states[:, 0, :, stack_number]
doc_obs = next_states[:, 1:, :, stack_number]
batch_size = next_q_values.get_shape().as_list()[0]
next_sarsa_q_list = []
for i in range(batch_size):
s, s_no_click = score_documents_tf(user_obs[i], doc_obs[i])
q = next_q_values[i]
slate = tf.expand_dims(next_actions[i], 1)
p_selected = compute_probs_tf(slate, s, s_no_click)
q_selected = tf.gather(q, slate)
next_sarsa_q_list.append(
tf.reduce_sum(input_tensor=p_selected * q_selected))
next_sarsa_q_values = tf.stack(next_sarsa_q_list)
return reward + gamma * next_sarsa_q_values * (
1. - tf.cast(terminals, tf.float32))
def true_fn(images):
if augment_entire_batch:
image_2 = images
mean_color = tf.reduce_mean(image_2, axis=[1, 2], keepdims=True)
print(mean_color.shape)
else:
image_1, image_2 = tf.unstack(images)
mean_color = tf.reduce_mean(image_2, axis=[0, 1], keepdims=True)
def body(var_img, mean_color):
x0 = tf.random.uniform([], 0, width, dtype=tf.int32)
y0 = tf.random.uniform([], 0, height, dtype=tf.int32)
dx = tf.random.uniform([], min_size, max_size, dtype=tf.int32)
dy = tf.random.uniform([], min_size, max_size, dtype=tf.int32)
x = tf.range(width)
x_mask = (x0 <= x) & (x < x0 + dx)
y = tf.range(height)
y_mask = (y0 <= y) & (y < y0 + dy)
mask = x_mask & y_mask[:, tf.newaxis]
mask = tf.cast(mask[:, :, tf.newaxis], image_2.dtype)
result = var_img * (1 - mask) + mean_color * mask
return result
# Perform at least one erase operation.
image_2 = body(image_2, mean_color)
# Perform additional erase operations.
for _ in range(max_operations - 1):
perform_erase = tf.less(tf.random.uniform([]),
probability_additional_operations)
image_2 = tf.cond(perform_erase, lambda: body(image_2, mean_color),
lambda: image_2)
if augment_entire_batch:
images = image_2
else:
images = tf.stack([image_1, image_2])
return images
def true_fn(images): image_1, image_2 = tf.unstack(images) image_1 = tf.image.random_brightness(image_1, max_delta) image_2 = tf.image.random_brightness(image_2, max_delta) return tf.stack([image_1, image_2])
def true_fn(images): image_1, image_2 = tf.unstack(images) image_1 = tf.image.random_contrast(image_1, min_bound, max_bound) image_2 = tf.image.random_contrast(image_2, min_bound, max_bound) return tf.stack([image_1, image_2])
def true_fn(images, flow, mask):
angle_radian = tf.random.uniform(
[], minval=-max_rotation, maxval=max_rotation,
dtype=tf.float32) * math.pi / 180.0
image_1, image_2 = tf.unstack(images)
image_2 = rotate(image_2, angle_radian, is_flow=False, mask=None)
images = tf.stack([image_1, image_2])
if not_empty_crop:
orig_height = tf.shape(images)[-3]
orig_width = tf.shape(images)[-2]
# introduce abbreviations for shorter notation
cos = tf.math.cos(angle_radian % math.pi)
sin = tf.math.sin(angle_radian % math.pi)
h = tf.cast(orig_height, tf.float32)
w = tf.cast(orig_width, tf.float32)
# compute required scale factor
scale = tf.cond(tf.math.less(angle_radian % math.pi, math.pi/2.0),
lambda: tf.math.maximum((w/h)*sin+cos, (h/w)*sin+cos),
lambda: tf.math.maximum((w/h)*sin-cos, (h/w)*sin-cos))
new_height = tf.math.floor(h / scale)
new_width = tf.math.floor(w / scale)
# crop image again to original size
offset_height = tf.cast((h-new_height)/2, tf.int32)
offset_width = tf.cast((w-new_width)/2, tf.int32)
images = tf.image.crop_to_bounding_box(
images,
offset_height=offset_height,
offset_width=offset_width,
target_height=tf.cast(new_height, tf.int32),
target_width=tf.cast(new_width, tf.int32))
if flow is not None:
# get current locations (with the origin in the image center)
positions = _positions_center_origin(orig_height, orig_width)
# compute augmented flow (multiply by mask to zero invalid flow locations)
cos = tf.math.cos(angle_radian)
sin = tf.math.sin(angle_radian)
rotation_matrix = tf.reshape([cos, sin, -sin, cos], [2, 2])
flow = (tf.linalg.matmul(
(positions + flow), rotation_matrix) - positions) * mask
if not_empty_crop:
# crop flow and mask again to original size
flow = tf.image.crop_to_bounding_box(
flow,
offset_height=offset_height,
offset_width=offset_width,
target_height=tf.cast(new_height, tf.int32),
target_width=tf.cast(new_width, tf.int32))
mask = tf.image.crop_to_bounding_box(
mask,
offset_height=offset_height,
offset_width=offset_width,
target_height=tf.cast(new_height, tf.int32),
target_width=tf.cast(new_width, tf.int32))
return images, flow, mask
def random_crop(images,
flow,
mask,
crop_height,
crop_width,
relative_offset,
probability_crop_offset):
"""Performs a random crop with the given height and width."""
# early return if crop_height or crop_width is not specified
if crop_height is None or crop_width is None:
return images, flow, mask
orig_height = tf.shape(images)[-3]
orig_width = tf.shape(images)[-2]
# check if crop size fits the image size
scale = 1.0
ratio = tf.cast(crop_height, tf.float32) / tf.cast(orig_height, tf.float32)
scale = tf.math.maximum(scale, ratio)
ratio = tf.cast(crop_width, tf.float32) / tf.cast(orig_width, tf.float32)
scale = tf.math.maximum(scale, ratio)
# compute minimum required hight
new_height = tf.cast(
tf.math.ceil(tf.cast(orig_height, tf.float32) * scale), tf.int32)
new_width = tf.cast(
tf.math.ceil(tf.cast(orig_width, tf.float32) * scale), tf.int32)
# perform resize (scales with 1 if not required)
images = smurf_utils.resize(images, new_height, new_width, is_flow=False)
# compute joint offset
max_offset_h = new_height - tf.cast(crop_height, dtype=tf.int32)
max_offset_w = new_width - tf.cast(crop_width, dtype=tf.int32)
joint_offset_h = tf.random.uniform([], maxval=max_offset_h+1, dtype=tf.int32)
joint_offset_w = tf.random.uniform([], maxval=max_offset_w+1, dtype=tf.int32)
# compute relative offset
min_relative_offset_h = tf.math.maximum(
joint_offset_h - relative_offset, 0)
max_relative_offset_h = tf.math.minimum(
joint_offset_h + relative_offset, max_offset_h)
min_relative_offset_w = tf.math.maximum(
joint_offset_w - relative_offset, 0)
max_relative_offset_w = tf.math.minimum(
joint_offset_w + relative_offset, max_offset_w)
relative_offset_h = tf.random.uniform(
[], minval=min_relative_offset_h, maxval=max_relative_offset_h+1,
dtype=tf.int32)
relative_offset_w = tf.random.uniform(
[], minval=min_relative_offset_w, maxval=max_relative_offset_w+1,
dtype=tf.int32)
set_crop_offset = tf.random.uniform([]) < probability_crop_offset
relative_offset_h = tf.cond(
set_crop_offset, lambda: relative_offset_h, lambda: joint_offset_h)
relative_offset_w = tf.cond(
set_crop_offset, lambda: relative_offset_w, lambda: joint_offset_w)
# crop both images
image_1, image_2 = tf.unstack(images)
image_1 = tf.image.crop_to_bounding_box(
image_1, offset_height=joint_offset_h, offset_width=joint_offset_w,
target_height=crop_height, target_width=crop_width)
image_2 = tf.image.crop_to_bounding_box(
image_2, offset_height=relative_offset_h, offset_width=relative_offset_w,
target_height=crop_height, target_width=crop_width)
images = tf.stack([image_1, image_2])
if flow is not None:
# perform resize (scales with 1 if not required)
flow, mask = smurf_utils.resize(
flow, new_height, new_width, is_flow=True, mask=mask)
# crop flow and mask
flow = tf.image.crop_to_bounding_box(
flow,
offset_height=joint_offset_h,
offset_width=joint_offset_w,
target_height=crop_height,
target_width=crop_width)
mask = tf.image.crop_to_bounding_box(
mask,
offset_height=joint_offset_h,
offset_width=joint_offset_w,
target_height=crop_height,
target_width=crop_width)
# correct flow for relative shift (/crop)
flow_delta = tf.stack(
[tf.cast(relative_offset_h - joint_offset_h, tf.float32),
tf.cast(relative_offset_w - joint_offset_w, tf.float32)])
flow = (flow - flow_delta) * mask
return images, flow, mask, joint_offset_h, joint_offset_w
def false_fn(scale_height, scale_width): return tf.stack((scale_height, scale_width), axis=0)
def true_fn(scale_height, scale_width): scale_height *= 2 ** tf.random.uniform([], -max_strech, max_strech) scale_width *= 2 ** tf.random.uniform([], -max_strech, max_strech) return tf.stack((scale_height, scale_width), axis=0)
def true_fn(images): image_1, image_2 = tf.unstack(images) image_1 = potential_asymmetric_augmentations(image_1) image_2 = potential_asymmetric_augmentations(image_2) return tf.stack([image_1, image_2])
def photometric_augmentation(images,
augment_color_swap=True,
augment_hue_shift=True,
augment_saturation=False,
augment_brightness=False,
augment_contrast=False,
augment_gaussian_noise=False,
augment_brightness_individual=False,
augment_contrast_individual=False,
max_delta_hue=0.5,
min_bound_saturation=0.8,
max_bound_saturation=1.2,
max_delta_brightness=0.1,
min_bound_contrast=0.8,
max_bound_contrast=1.2,
min_bound_gaussian_noise=0.0,
max_bound_gaussian_noise=0.02,
max_delta_brightness_individual=0.02,
min_bound_contrast_individual=0.95,
max_bound_contrast_individual=1.05):
"""Applies photometric augmentations to an image pair."""
# Randomly permute colors by rolling and reversing.
# This covers all permutations.
if augment_color_swap:
r = tf.random.uniform([], maxval=3, dtype=tf.int32)
images = tf.roll(images, r, axis=-1)
r = tf.equal(tf.random.uniform([], maxval=2, dtype=tf.int32), 1)
images = tf.cond(pred=r,
true_fn=lambda: tf.reverse(images, axis=[-1]),
false_fn=lambda: images)
if augment_hue_shift:
images = tf.image.random_hue(images, max_delta_hue)
if augment_saturation:
images = tf.image.random_saturation(
images, min_bound_saturation, max_bound_saturation)
if augment_brightness:
images = tf.image.random_brightness(images, max_delta_brightness)
if augment_contrast:
images = tf.image.random_contrast(
images, min_bound_contrast, max_bound_contrast)
if augment_gaussian_noise:
sigma = tf.random.uniform([],
minval=min_bound_gaussian_noise,
maxval=max_bound_gaussian_noise,
dtype=tf.float32)
noise = tf.random.normal(
tf.shape(input=images), stddev=sigma, dtype=tf.float32)
images = images + noise
# perform relative photometric augmentation (individually per image)
image_1, image_2 = tf.unstack(images)
if augment_brightness_individual:
image_1 = tf.image.random_contrast(
image_1, min_bound_contrast_individual, max_bound_contrast_individual)
image_2 = tf.image.random_contrast(
image_2, min_bound_contrast_individual, max_bound_contrast_individual)
if augment_contrast_individual:
image_1 = tf.image.random_brightness(
image_1, max_delta_brightness_individual)
image_2 = tf.image.random_brightness(
image_2, max_delta_brightness_individual)
# crop values to ensure values in [0,1] (some augmentations can violate this)
image_1 = tf.clip_by_value(image_1, 0.0, 1.0)
image_2 = tf.clip_by_value(image_2, 0.0, 1.0)
return tf.stack([image_1, image_2])
def _non_nan_mean(tensor_list): """Calculates the mean of a list of tensors while ignoring nans.""" tensor = tf.stack(tensor_list) not_nan = tf.logical_not(tf.math.is_nan(tensor)) return tf.reduce_mean(tf.boolean_mask(tensor, not_nan))
def classification_loss_using_mask_iou_func(embeddings,
logits,
instance_ids,
class_labels,
num_samples,
valid_mask=None,
max_instance_id=None,
similarity_strategy='dotproduct',
is_balanced=True):
"""Classification loss using mask iou.
Args:
embeddings: A tf.float32 tensor of size [batch_size, n, f].
logits: A tf.float32 tensor of size [batch_size, n, num_classes]. It is
assumed that background is class 0.
instance_ids: A tf.int32 tensor of size [batch_size, n].
class_labels: A tf.int32 tensor of size [batch_size, n]. It is assumed
that the background voxels are assigned to class 0.
num_samples: An int determining the number of samples.
valid_mask: A tf.bool tensor of size [batch_size, n] that is True when an
element is valid and False if it needs to be ignored. By default the value
is None which means it is not applied.
max_instance_id: If set, instance ids larger than that value will be
ignored. If not set, it will be computed from instance_ids tensor.
similarity_strategy: Defines the method for computing similarity between
embedding vectors. Possible values are 'dotproduct' and
'distance'.
is_balanced: If True, the per-voxel losses are re-weighted to have equal
total weight for foreground vs. background voxels.
Returns:
A tf.float32 scalar loss tensor.
"""
batch_size = embeddings.get_shape().as_list()[0]
if batch_size is None:
raise ValueError('Unknown batch size at graph construction time.')
if max_instance_id is None:
max_instance_id = tf.reduce_max(instance_ids)
class_labels = tf.reshape(class_labels, [batch_size, -1, 1])
sampled_embeddings, sampled_instance_ids, sampled_indices = (
sampling_utils.balanced_sample(features=embeddings,
instance_ids=instance_ids,
num_samples=num_samples,
valid_mask=valid_mask,
max_instance_id=max_instance_id))
losses = []
for i in range(batch_size):
embeddings_i = embeddings[i, :, :]
instance_ids_i = instance_ids[i, :]
class_labels_i = class_labels[i, :, :]
logits_i = logits[i, :]
sampled_embeddings_i = sampled_embeddings[i, :, :]
sampled_instance_ids_i = sampled_instance_ids[i, :]
sampled_indices_i = sampled_indices[i, :]
sampled_class_labels_i = tf.gather(class_labels_i, sampled_indices_i)
sampled_logits_i = tf.gather(logits_i, sampled_indices_i)
if valid_mask is not None:
valid_mask_i = valid_mask[i]
embeddings_i = tf.boolean_mask(embeddings_i, valid_mask_i)
instance_ids_i = tf.boolean_mask(instance_ids_i, valid_mask_i)
loss_i = classification_loss_using_mask_iou_func_unbatched(
embeddings=embeddings_i,
instance_ids=instance_ids_i,
sampled_embeddings=sampled_embeddings_i,
sampled_instance_ids=sampled_instance_ids_i,
sampled_class_labels=sampled_class_labels_i,
sampled_logits=sampled_logits_i,
similarity_strategy=similarity_strategy,
is_balanced=is_balanced)
losses.append(loss_i)
return tf.math.reduce_mean(tf.stack(losses))
def prepare_lidar_images_and_correspondences(
inputs,
resized_image_height,
resized_image_width,
camera_names=('front', 'front_left', 'front_right', 'side_left',
'side_right'),
lidar_names=('top', 'front', 'side_left', 'side_right', 'rear')):
"""Integrates and returns the lidars, cameras and their correspondences.
Args:
inputs: A dictionary containing the images and point / pixel
correspondences.
resized_image_height: Target height of the images.
resized_image_width: Target width of the images.
camera_names: List of cameras to include images from.
lidar_names: List of lidars to include point clouds from.
Returns:
A tf.float32 tensor of size [num_points, 3] containing point positions.
A tf.float32 tensor of size [num_points, 1] containing point intensities.
A tf.float32 tensor of size [num_points, 1] containing point elongations.
A tf.float32 tensor of size [num_points, 3] containing point normals.
A tf.float32 tensor of size [num_images, resized_image_height,
resized_image_width, 3].
A tf.int32 tensor of size [num_images, num_points, 2].
Raises:
ValueError: If camera_names or lidar_names are empty lists.
"""
if not camera_names:
raise ValueError('camera_names should contain at least one name.')
if not lidar_names:
raise ValueError('lidar_names should contain at least one name.')
(points_position, points_intensity, points_elongation, points_normal,
points_in_image_frame_yx, points_in_image_frame_id) = _prepare_lidar_points(
inputs=inputs, lidar_names=lidar_names)
images = []
points_in_image_frame = []
for camera_name in camera_names:
image_key = ('cameras/%s/image' % camera_name)
image_height = tf.shape(inputs[image_key])[0]
image_width = tf.shape(inputs[image_key])[1]
height_ratio = tf.cast(
resized_image_height, dtype=tf.float32) / tf.cast(
image_height, dtype=tf.float32)
width_ratio = tf.cast(
resized_image_width, dtype=tf.float32) / tf.cast(
image_width, dtype=tf.float32)
if tf.executing_eagerly():
resize_method = tf.image.ResizeMethod.NEAREST_NEIGHBOR
else:
resize_method = tf.image.ResizeMethod.BILINEAR
if inputs[image_key].dtype in [
tf.int8, tf.uint8, tf.int16, tf.uint16, tf.int32, tf.int64
]:
resize_method = tf.image.ResizeMethod.NEAREST_NEIGHBOR
images.append(
tf.image.resize(
images=inputs[image_key],
size=[resized_image_height, resized_image_width],
method=resize_method,
antialias=True))
camera_id = tf.cast(inputs[('cameras/%s/id' % camera_name)], dtype=tf.int32)
valid_points = tf.equal(points_in_image_frame_id, camera_id)
valid_points = tf.tile(valid_points, [1, 2])
point_coords = tf.cast(
tf.cast(points_in_image_frame_yx, dtype=tf.float32) *
tf.stack([height_ratio, width_ratio]),
dtype=tf.int32)
points_in_image_frame_camera = tf.where(
valid_points, point_coords, -tf.ones_like(valid_points, dtype=tf.int32))
points_in_image_frame.append(points_in_image_frame_camera)
num_images = len(images)
images = tf.stack(images, axis=0)
images.set_shape([num_images, resized_image_height, resized_image_width, 3])
points_in_image_frame = tf.stack(points_in_image_frame, axis=0)
return {
'points_position': points_position,
'points_intensity': points_intensity,
'points_elongation': points_elongation,
'points_normal': points_normal,
'view_images': {'rgb_view': images},
'view_indices_2d': {'rgb_view': points_in_image_frame}
}
def flip_points(points, x_rotate, y_rotate):
return points * tf.stack((x_rotate, y_rotate, 1), axis=0)