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

Define kernel size which is automatically reduced for small input. If the shape of the input images is unknown at graph construction time this fu

Return angle between vectors. If directed is False, the input vectors are interpreted as undirected axes, i.e. the maximum angle is pi/2.

Bottleneck residual unit variant with BN after convolutions. This is the original residual unit proposed in [1]. See Fig. 1(a) of [2] for its def

Return matrix to obtain normalized device coordinates from frustum. The frustum bounds are axis-aligned along x (left, right), y (bottom, top

Return transformation matrix from sequence of transformations. This is the inverse of the decompose_matrix function. Sequence of transformat

Return concatenation of series of transformation matrices. >>> M = numpy.random.rand(16).reshape((4, 4)) - 0.5 >>> numpy.allclose(M, concaten

Configures the optimizer used for training. Args: learning_rate: A scalar or `Tensor` learning rate. Returns: An instance of an

Return state of constrain to axis mode.

Strided 2-D convolution with 'SAME' padding. When stride > 1, then we do explicit zero-padding, followed by conv2d with 'VALID' padding. Note

Return sequence of transformations from transformation matrix. matrix : array_like Non-degenerative homogeneous transformation matrix

Distort the color of a Tensor image. Each color distortion is non-commutative and thus ordering of the color ops matters. Ideally we would ran

Set initial cursor window coordinates and pick constrain-axis.

Update current cursor window coordinates.

Return Euler angles from quaternion for specified axis sequence. >>> angles = euler_from_quaternion([0.99810947, 0.06146124, 0, 0]) >>> numpy

:param p1: [x1, y1] :param p2: [x2, y2] :return: line of form [a, b, c] from equation of form ax+by+c=0

Returns a function run by the chief worker to warm-start the training. Note that the init_fn is only run when initializing the model during the v

Params: ------- p1, p2: belong to a line (3D) q: another point (3D) Returns: -------- a 'point' on the line containing p

Params: ------- p1, p2: belong to a line (assume that is on image) (3D) sphere_centre, sphere_radius: params of a sphere located on t

Returns a list of variables to train. Returns: A list of variables to train by the optimizer.

Return 4x4 identity/unit matrix. >>> I = identity_matrix() >>> numpy.allclose(I, numpy.dot(I, I)) True >>> numpy.sum(I), numpy.trace(

Defines the default arg scope for inception models. Args: weight_decay: The weight decay to use for regularizing the model. use_batch_norm:

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

Creates the Inception V4 model. Args: inputs: a 4-D tensor of size [batch_size, height, width, 3]. num_classes: number of predicted classes

Return inverse of square transformation matrix. >>> M0 = random_rotation_matrix() >>> M1 = inverse_matrix(M0.T) >>> numpy.allclose(M1, nu

Return True if two quaternions are equal.

Return True if two matrices perform same transformation. >>> is_same_transform(numpy.identity(4), numpy.identity(4)) True >>> is_same_tra

Return homogeneous rotation matrix.

Mobilenet v1 model for classification. Args: inputs: a tensor of shape [batch_size, height, width, channels]. num_classes: number of predic

Defines the default MobilenetV1 arg scope. Args: is_training: Whether or not we're training the model. If this is set to None, the parame

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

Continue rotation in direction of last drag.

Return orthogonalization matrix for crystallographic cell coordinates. Angles are expected in degrees. The de-orthogonalization matrix is th

Place Arcball, e.g. when window size changes. center : sequence[2] Window coordinates of trackball center. radius : float

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

Return projection plane and perspective point from projection matrix. Return values are same as arguments for projection_matrix function: poi

Return matrix to project onto plane defined by point and normal. Using either perspective point, projection direction, or none of both. If p

Return quaternion for rotation about axis. >>> q = quaternion_about_axis(0.123, [1, 0, 0]) >>> numpy.allclose(q, [0.99810947, 0.06146124, 0,

Return conjugate of quaternion. >>> q0 = random_quaternion() >>> q1 = quaternion_conjugate(q0) >>> q1[0] == q0[0] and all(q1[1:] == -q0[1

Return quaternion from Euler angles and axis sequence. ai, aj, ak : Euler's roll, pitch and yaw angles axes : One of 24 axis sequences as str

Return imaginary part of quaternion. >>> quaternion_imag([3, 0, 1, 2]) array([ 0., 1., 2.])

Return inverse of quaternion. >>> q0 = random_quaternion() >>> q1 = quaternion_inverse(q0) >>> numpy.allclose(quaternion_multiply(q0, q1)

Return real part of quaternion. >>> quaternion_real([3, 0, 1, 2]) 3.0

Return uniform random rotation matrix. rand: array like Three independent random variables that are uniformly distributed between

Return array of random doubles in the half-open interval [0.0, 1.0). >>> v = random_vector(10000) >>> numpy.all(v >= 0) and numpy.all(v < 1)

Return mirror plane point and normal vector from reflection matrix. >>> v0 = numpy.random.random(3) - 0.5 >>> v1 = numpy.random.random(3) - 0

Return matrix to mirror at plane defined by point and normal vector. >>> v0 = numpy.random.random(4) - 0.5 >>> v0[3] = 1. >>> v1 = numpy.

Defines the default ResNet arg scope. TODO(gpapan): The batch-normalization related default values above are appropriate for use in conjunction

ResNet-101 model of [1]. See resnet_v1() for arg and return description.

ResNet-152 model of [1]. See resnet_v1() for arg and return description.

ResNet-200 model of [2]. See resnet_v1() for arg and return description.

ResNet-50 model of [1]. See resnet_v1() for arg and return description.

Rotate a point counterclockwise by a given (+ve) angle around a given origin. The angle should be given in radians.

Given a rectangle of size wxh that has been rotated by 'angle' (in radians), computes the width and height of the largest possible axis-a

Return rotation angle and axis from rotation matrix. >>> angle = (random.random() - 0.5) * (2*math.pi) >>> direc = numpy.random.random(3) - 0

:param input_line: line of form [a, b, c] from equation of form ax+by+c=0 :param orig_dims: tuple of (width, height) of original dimensions

Return scaling factor, origin and direction from scaling matrix. >>> factor = random.random() * 10 - 5 >>> origin = numpy.random.random(3) -

Return matrix to scale by factor around origin in direction. Use factor -1 for point symmetry. >>> v = (numpy.random.rand(4, 5) - 0.5) * 20

Set axes to constrain rotations.

Return shear angle, direction and plane from shear matrix. >>> angle = (random.random() - 0.5) * 4*math.pi >>> direct = numpy.random.random(3

Return matrix to shear by angle along direction vector on shear plane. The shear plane is defined by a point and normal vector. The direction

Stacks ResNet `Blocks` and controls output feature density. First, this function creates scopes for the ResNet in the form of 'block_name/unit_1'

Return matrix to transform given 3D point set into second point set. v0 and v1 are shape (3, \*) or (4, \*) arrays of at least 3 points. The

Return translation vector from translation matrix. >>> v0 = numpy.random.random(3) - 0.5 >>> v1 = translation_from_matrix(translation_matrix(

Return matrix to translate by direction vector. >>> v = numpy.random.random(3) - 0.5 >>> numpy.allclose(v, translation_matrix(v)[:3, 3])

Return vector perpendicular to vectors. >>> v = vector_product([2, 0, 0], [0, 3, 0]) >>> numpy.allclose(v, [0, 0, 6]) True >>> v0 = [

Oxford Net VGG 16-Layers version D Example. Note: All the fully_connected layers have been transformed to conv2d layers. To use in classifi

Oxford Net VGG 19-Layers version E Example. Note: All the fully_connected layers have been transformed to conv2d layers. To use in classifi