Modified from pytorch3d and add `model_type` in __class__.__init__.

Concatenate two human_data. All keys will be kept it the returned human_data. If either value from human_data_0 or human_data_1 matches

Update a new data sample.

Convert values in numpy.ndarray type to torch.Tensor, and move Tensors to the target device. All keys will exist in the returned dict.

Build loss.

Create new camera class which contains each input camera N times. Args: N: number of new copies of each camera. Returns:

Build body_models.

Load data from npz_path and update them to self. Args: npz_path (str): Path to a dumped npz file.

Convert keypoints following the mapping correspondence between src and dst keypoints definition. Supported conventions by now: agora, coco, smplx,

Returns parameters. Compatible with mmcv's build_parameters() Returns: opt_params: a list of body model parameters for optimizati

Convert rotation matrixs to axis angles. Args: matrix (Union[torch.Tensor, numpy.ndarray]): input shape should be (..., 3

Set self[key] to value. Only be called when using human_data[key] = val. Methods like update won't call __setitem__. In keypoints_comp

Get keypoint index from specified convention with keypoint name. Args: name (str): keypoint name convention (str): data type from

Get part keypoints indices from specified part and convention. Args: part (str): part to search from convention (str): data type

Check whether the suffix of the path is allowed. Args: path_str (str): Path to check. allowed_suffix (List[str], opti

Convert 6D rotation representation to 3x3 rotation matrix. Based on Zhou et al., "On the Continuity of Rotation Representations in Neural Net

Check output folder or file. Args: output_path (str): could be folder or file. allowed_suffix (List[str], optional):

Perform orthographic projection of 3D points using the camera parameters, return projected 2D points in image plane. Notes: batch siz

Change the device of the data inplace. Example: >>> self = SamplingResult.random() >>> print(f'self = {self.to(None)}

Check input folder or file. Args: input_path (str): input folder or file path. allowed_suffix (List[str], optional):

Get raw value from the dict. It acts the same as dict.__getitem__(k). Args: key (_KT): Key in dict.

A helper function to ensure body model parameters have the same batch size as the input keypoints. Args: init_param: inpu

Returns the image size, if provided, expected in the form of (height, width) The image size is used for conversion of projected points

Set requires_grad and collect parameters for optimization. Args: fit_param: whether to optimize this body model parameter

Boolean of whether in NDC.

dict: a dictionary of info about the object

Convert axis-angle representation to rotation matrix. Args: theta: size = [B, 3] Returns: Rotation matrix corresponding to th

Get mean value of params.

Get a parameter by key. Args: key (str): Name of the parameter. Raises: KeyError: key not in

Convert axis_angle to rotation matrixs. Args: axis_angle (Union[torch.Tensor, numpy.ndarray]): input shape should be

Check whether a file or a directory exists at the expected path. Args: path_str (str): Path to check. path_type (Lite

Dump keys and items to an npz file. Args: npz_path (str): Path to a dumped npz file. overwrite (bool,

Get a numpy of colors of shape (N, 3).

Performs Linear Blend Skinning with the given shape and pose parameters. Parameters ---------- betas : torch.tensor BxNB The tens

Normalize the tensor or array and convert dtype.

Scatter function. Args: inputs (Tensor): Input Tensor. kwargs (dict): Args for ``mmcv.parallel.scatte

Calculates the 3D joint locations from the vertices. Parameters ---------- J_regressor : torch.tensor JxV The regressor array tha

Transform the bbox format from x1y1x2y2 to xywh. Args: bbox_xyxy (np.ndarray): Bounding boxes (with scores), shaped (n, 4) or

SMPL model constructor. Parameters ---------- model_path: str The path to the folder or to the file where the mod

Get number of keypoints of specified convention. Args: convention (str): data type from keypoints_factory. keypoints_factory (dic

Apply function to a list of arguments. Note: This function applies the ``func`` to multiple inputs and map the multiple outputs o

Whether to skip loss computation. If loss is None, it will directly skip the loss to avoid RuntimeError. If loss is not None, the table

Apply an affine transformation to the points. Args: pt (np.ndarray): a 2 dimensional point to be transformed trans_mat (np.ndarra

Convert the intrinsic matrix K and extrinsic matrix [R|T] from source convention to destination convention. Args: K (Union[torch.Tens

Get 3d keypoints and visibility mask Args: keypoints (np.ndarray): 2d (NxKx3) or 3d (NxKx4) keypoints with visibility

nn.Module: the normalization layer named "norm1"

nn.Module: normalization layer after the second convolution layer

Create random AssignResult for tests or debugging. Args: num_preds: number of predicted boxes num_gts: number of true

Set a matrix-type parameter to mat_numpy. Args: mat_key (str): Key of the target matrix. in_mat or rotation_mat.

Set a parameter to value. Args: key (str): Name of the parameter. value (object): New

Build backbone.

Calculate the mean per-joint position error (MPJPE) and the error after rigid alignment with the ground truth (PA-MPJPE). batch_size: N nu

Flip human joints horizontally. Note: num_keypoints: K num_dimension: D Args: keypoints (np.ndarray([K, D])): Coordin

Parse results.

Parse results.

Rotate SMPL pose parameters. SMPL (https://smpl.is.tue.mpg.de/) is a 3D human model. Args: pose (np.ndarray([72])): SMPL pose par

Convert xywh coordinates to center and scale. Args: bbox_xywh (numpy.ndarray): the height of the bbox_xywh aspect_ratio (int, optional):

Build cameras.

Build head.