OpenCV intrinsics to OpenGL perspective matrix Args: intrinsics (torch.Tensor): [3, 3] OpenCV intrinsics matrix near (float)

OpenCV intrinsics to OpenGL perspective matrix Args: intrinsics (torch.Tensor): [3, 3] OpenCV intrinsics matrix near (float)

Check if the mask touches the image border within a given thickness.

Efficient occlusion detection using depth map and internal/external edges.

Load 3DB mesh and convert from OpenGL to PyTorch3D coordinates.

Load focal length from JSON file.

Load mask image as numpy array.

Create a mesh from pointmap and image. Returns: faces, vertices, vertex_colors, vertex_uvs: Mesh components

Normalize pointmap using Scale-Shift Invariant (SSI) normalization. Args: pointmap: Pointmap tensor of shape (H, W, 3) or (3, H,

Parametrize a mesh to a texture space, using xatlas. Args: vertices (np.array): Vertices of the mesh. Shape (V, 3). faces (n

Patchify a tensor. Args: x (torch.Tensor): (N, C, *spatial) tensor patch_size (int): Patch size

3D pixel shuffle.

Args: pointmap: (H, W, 3) tensor of point coordinates image: (H, W, 3) tensor of image

Args: pointmap: (H, W, 3) tensor of point coordinates image: (H, W, 3) tensor of image

Postprocess a mesh by simplifying, removing invisible faces, and removing isolated pieces. Args: vertices (np.array): Vertices of th

Prepare RGB ground truth image for supervision by handling format conversion and resizing. Args: rgb_gt: Ground truth RGB image [3,

Complete pipeline for aligning 3DB mesh to MoGe scale.

Convert PyTorch3D quaternions (w,x,y,z) to SciPy-style Euler angles with sequence 'XYZ' (extrinsic, radians). Works with batch dims. Arg

Remove the background from an image using a mask. For pointmaps, sets background regions to NaN. This function follows the standard tran

Remove small disconnected regions (floating points) from the mask. Keeps all regions with area >= min_area.

Render the scene. Background tensor (bg_color) must be on GPU!

Render the scene. Background tensor (bg_color) must be on GPU!

Create image and mask transforms based on configuration. Returns: dict: {"img_transform": img_transform, "mask_transform": mask_tran

Blur the input image by applying upsample(downsample(x)). Args: image (torch.Tensor): Image tensor of shape (C, H, W), float32, with

Restore Gaussian to its original state using backup data.

Parameters: - image (Image): The input image to be processed. - seed (int, optional): The random seed for reproducibility. De

Run ICP alignment on Gaussian positions. Similar to run_ICP but for GS.

Manual alignment between Gaussian positions and point cloud. Similar to run_alignment but for GS.

Args: gaussian: Gaussian splatting object center: Center point for rotation renderer: GS renderer intrinsics: Cam

Apply serialized scaled dot product self attention to a sparse tensor. Args: qkv (SparseTensor): [N, *, 3, H, C] sparse tensor conta

Unbind a sparse tensor along a dimension. Args: input (SparseTensor): Sparse tensor to unbind. dim (int): Dimension to unbin

Apply windowed scaled dot product self attention to a sparse tensor. Args: qkv (SparseTensor): [N, *, 3, H, C] sparse tensor contain

Create sinusoidal timestep embeddings. Args: t: a 1-D Tensor of N indices, one per batch element. These

Create sinusoidal timestep embeddings. Args: t: a 1-D Tensor of N indices, one per batch element. These