input: pos output: mask

Returns the standard names for spherical harmonic coefficients. This method generates a list of names for the 15 SH coefficients use

Random gaussians

Pytorch implementation of `gsplat.cuda._wrapper.isect_offset_encode()`. .. note:: This is a minimal implementation of the fully fused ve

Pytorch implementation of `gsplat.cuda._wrapper.isect_tiles()`. .. note:: This is a minimal implementation of the fully fused version, w

Loads metadata from a JSON file in the specified directory.

Default YUV loading.

Pad the attribute sequences to make them square. Args: splats_videos (Dict[str, Tensor]): Dictionary of attribute sequences, where

Pad the splats to make them square Args: splats (Dict[str, Tensor]): Dictionary of splats with keys as attributes and values as te

Quantizes a tensor to a specified bit depth using per-channel min-max normalization.

Convert quaternion and scale to a 3x3 matrix (R * S).

A version of rasterization() that utilies on PyTorch's autograd. .. note:: This function still relies on gsplat's CUDA backend for some c

Pytorch implementation of `gsplat.cuda._wrapper.rasterize_to_pixels_2dgs()`. This function rasterizes 2D Gaussians to pixels in a Pytorch-friendl

Reorganize a list of splat dictionaries into a sequence of attributes. Args: splats_list (List[Dict]): List of splat dictionaries,

Save lightweight compression metadata and return it.

Saves metadata to a JSON file in the specified directory.

Organize the list of splats into several sequences of attributs Args: splats_list (List[Dict]): List of splat dictionaries, each

Pytorch implementation of `gsplat.cuda._wrapper.spherical_harmonics()`.

Convert a list of splat dictionaries into a sequence of attributes. Args: splats_list (List[Dict]): List of splat dictionaries, ea

Write minimal metadata required for decoding.

Alpah compositing of 2D Gaussians in Pure Pytorch. This function performs alpha compositing for Gaussians based on the pair of indices {gauss

Alpha compositing for 2DGS. .. warning:: This function requires the nerfacc package to be installed. Please install it using the followin

Draw a red rectangle highlighting the region and place a x2 zoomed inset of the region at the bottom-right corner of the image. Args:

Convert single channel to a color img. Args: img (torch.Tensor): (..., 1) float32 single channel image. colormap (str): Colormap

New pose using average position, z-axis, and up vector of input poses.

Instantiate or update an object using a registry mapping.

Convert total bytes across a sequence to kbps at given fps and frame_num. kbps = (size_bytes / frame_num) * 8 * fps / 1024

Convert total bytes across a sequence to kbps at given fps and frame_num. kbps = (size_bytes / frame_num) * 8 * fps / 1024

Calculate average metrics across all scenes. Args: scene_metrics: Dictionary containing metrics for each scene Retu

Calculate the total size of all files in a directory in MB. Args: directory_path: Path to the directory to calculate size for

Calculate PSNR between two images.

Best-effort coercion to keep tuple/list shapes consistent.

Compare rendering statistics between two sets of results. Args: stats1 (Dict): First set of rendering statistics

Compare rendering statistics between two sets of results. Args: stats1 (Dict): First set of rendering statistics

Run compression Args: compress_dir (str): directory to save compressed files splats (Dict[str, Tensor]): Gaussian spl

Entry for running video anchor compression.

Expand plotting bounds slightly around observed values for aesthetics. Parameters ---------- values : List[float] Sampled values