:param vertices: [batch size, number of vertices, 3] :param faces: [batch size, number of faces, 3] :return: [batch size, number of face

Calculates landmarks by barycentric interpolation Parameters ---------- vertices: torch.tensor BxVx3, dtype = torch.float32

landmarks_interpolate. :param landmarks: List, the raw landmark (in-place)

-- rendering shape with detail normal map

Calculates the rotation matrices for a batch of rotation vectors Parameters ---------- rot_vecs: torch.tensor Nx3

Calculates the rotation matrices for a batch of rotation vectors Parameters ---------- rot_vecs: torch.tensor Nx3

function adapted from https://github.com/mpc001/Visual_Speech_Recognition_for_Multiple_Languages

Draw 68 key points Args: image: the input image kpt: (68, 3).

:param vertices: [batch size, number of vertices, 3] :param faces: [batch size, number of faces, 3] :return: [batch size, number of verti

Calculates landmarks by barycentric interpolation Parameters ---------- vertices: torch.tensor BxVx3, dtype = torch.float32

Selects the face contour depending on the reletive position of the head Input: vertices: N X num_of_vertices

sh_coeff: [bz, 9, 3]

normals: [bz, nv, 3] lights: [bz, nlight, 6] returns: shading: [bz, nv, 3]

normals: [bz, nv, 3] lights: [bz, nlight, 6] returns: shading: [bz, nv, 3]

Computes the l1 loss between the ground truth keypoints and the predicted keypoints Inputs: kp_gt : N x K x 3 kp_pred: N x K x 2

orthgraphic projection X: 3d vertices, [bz, n_point, 3] camera: scale and translation, [bz, 3], [scale, tx, ty]

3x3 convolution with padding

3x3 convolution with padding

3x3 convolution with padding

Prints the person's name and age. If the argument 'additional' is passed, then it is appended after the main info. Paramete

NOW validation

r"""Function that returns Gaussian filter coefficients. Args: kernel_size (int): filter size. It should be odd and positive. sigm

Ref: https://github.com/facebookresearch/pytorch3d/blob/25c065e9dafa90163e7cec873dbb324a637c68b7/pytorch3d/io/obj_io.py Load a mesh from a file-l

Predicts the persons head pose and returning it in euler angles. Parameters ---------- img : array Face

Convert quaternion vector to angle axis of rotation. TODO: CORRECT Adapted from ceres C++ library: ceres-solver/include/ceres/rotation.h Arg

Convert quaternion coefficients to rotation matrix. Args: quat: size = [B, 4] 4 <===>(w, x, y, z) Returns: Rotation matrix cor

Render the mesh with the given vertices. Transformed vertices includes vertices in NDC space. Note that due to this custom implementation of

Initialize the Resnet block A resnet block is a conv block with skip connections We construct a conv block with build_conv_block func

r"""Creates a binary kernel to extract the patches. If the window size is HxW will create a (H*W)xHxW kernel.

Selects the face contour depending on the reletive position of the head Input: vertices: N X num_of_vertices

vertices: [bz, nv, 3] lights: [bz, nlight, 6] returns: shading: [bz, nv, 3]

Convert an angle axis to a quaternion. Adapted from ceres C++ library: ceres-solver/include/ceres/rotation.h Args: angle_axis (torch

Applies a batch of rigid transformations to the joints Parameters ---------- rot_mats : torch.tensor BxNx3x3 Tensor of rotat