Helper function for creating a resnet_v1 bottleneck block. Args: scope: The scope of the block. base_depth: The depth of the bottleneck lay

Return length, i.e. Euclidean norm, of ndarray along axis. >>> v = numpy.random.random(3) >>> n = vector_norm(v) >>> numpy.allclose(n, nu

Return ndarray normalized by length, i.e. Euclidean norm, along axis. >>> v0 = numpy.random.random(3) >>> v1 = unit_vector(v0) >>> numpy.

Return matrix to rotate about axis defined by point and direction. >>> R = rotation_matrix(math.pi/2, [0, 0, 1], [1, 0, 0]) >>> numpy.allclos

Notes: ------ - [min_val, max_val) - if exclude is a scalar value, then that value is not chosen during randomization

Return homogeneous rotation matrix from quaternion. >>> M = quaternion_matrix([0.99810947, 0.06146124, 0, 0]) >>> numpy.allclose(M, rotation_

Generator for v1 ResNet models. This function generates a family of ResNet v1 models. See the resnet_v1_*() methods for specific model instantiat

Subtracts the given means from each image channel. For example: means = [123.68, 116.779, 103.939] image = _mean_image_subtraction(image, m

Return sphere point perpendicular to axis.

Params: ------- image_coord: a point on image (3D) sphere_centre, sphere_radius: params of a sphere located on the image. sphere_cent

Resize images preserving the original aspect ratio. Args: image: A 3-D image `Tensor`. smallest_side: A python integer or scalar `Tensor` i

Crops the given image using the provided offsets and sizes. Note that the method doesn't assume we know the input image size but it does assume w

Pads the input along the spatial dimensions independently of input size. Pads the input such that if it was used in a convolution with 'VALID' padd

Return unit sphere coordinates from window coordinates.

Parameters ---------- sphere_centre sphere_radius p1 p2 Returns ------- returns points with respect to sphere c

Performs central crops of the given image list. Args: image_list: a list of image tensors of the same dimension but possibly varying chan

Try import all public attributes from module into global namespace. Existing attributes with name clashes are renamed with prefix. Attributes

Crops the given image using the provided offsets and sizes. Note that the method doesn't assume we know the input image size but it does assume w

Crops the given list of images. The function applies the same crop to each image in the list. This can be effectively applied when there are mult

Computes new shape with the smallest side equal to `smallest_side`. Computes new shape with the smallest side equal to `smallest_side` while pres

Plot a line from slope and intercept

Return affine transform matrix to register two point sets. v0 and v1 are shape (ndims, \*) arrays of at least ndims non-homogeneous coordinat

Return axis, which arc is nearest to point.

Builds Inception-A block for Inception v4 network.

Builds Inception-B block for Inception v4 network.

Builds Inception-C block for Inception v4 network.

Builds Reduction-A block for Inception v4 network.

Builds Reduction-B block for Inception v4 network.

Return Euler angles from rotation matrix for specified axis sequence. axes : One of 24 axis sequences as string or encoded tuple Note that m

Return homogeneous rotation matrix from Euler angles and axis sequence. ai, aj, ak : Euler's roll, pitch and yaw angles axes : One of 24 axis

Set the shape to 3 dimensional if we don't know anything else. Args: image: original image size result: flipped or transformed image

:param horizon_vectorform: line of form [a, b, c] from equation of form ax+by+c=0 :return: slope and intercept of input line

fov: field of view (degrees) pitch: positive means looking up (degrees) yaw: positive means looking right (degrees) roll: positive me

Defines the Inception V1 base architecture. This architecture is defined in: Going deeper with convolutions Christian Szegedy, Wei Liu, Yan

Creates the Inception V4 network up to the given final endpoint. Args: inputs: a 4-D tensor of size [batch_size, height, width, 3]. final_e

Mobilenet v1. Constructs a Mobilenet v1 network from inputs to the given final endpoint. Args: inputs: a tensor of shape [batch_size, height

Input must be 2 dimensional. Softmax is applied separately on each row

Input must be 2 dimensional. Softmax is applied separately on each row

Input must be 2 dimensional. Softmax is applied separately on each row

Preprocesses the given image for evaluation. Args: image: A `Tensor` representing an image of arbitrary size. output_height: The height of

Preprocesses the given image for training. Note that the actual resizing scale is sampled from [`resize_size_min`, `resize_size_max`]. Args:

Return quaternion from rotation matrix. If isprecise is True, the input matrix is assumed to be a precise rotation matrix and a faster algori

Return multiplication of two quaternions. >>> q = quaternion_multiply([4, 1, -2, 3], [8, -5, 6, 7]) >>> numpy.allclose(q, [28, -44, -14, 48])

Return spherical linear interpolation between two quaternions. >>> q0 = random_quaternion() >>> q1 = random_quaternion() >>> q = quaterni

Return uniform random unit quaternion. rand: array like or None Three independent random variables that are uniformly distributed

Subsamples the input along the spatial dimensions. Args: inputs: A `Tensor` of size [batch, height_in, width_in, channels]. factor: The sub

Convert a CARLA raw image to a BGRA numpy array.

Convert a CARLA raw image to a RGB numpy array.

Initialize virtual trackball control. initial : quaternion or rotation matrix