SYNTOLOGY HomeExplorerAtlasCodeMethodologyAboutDevelopersFeedPricing
Paper · 2312.00732 · 2023

Gaussian Grouping: Segment and Edit Anything in 3D Scenes

arXiv · PDF · Open in the Atlas

Code that ran

9 of 13 samples (functions and classes) lifted from this paper's linked repositories executed on a synthesized input in Syntology's jail (0 of them constructed an object rather than computing a result; 1 honoured the contract we drafted, 0 violated it and 0 have no contract verdict; 8 failures of our instrument; 4 no recorded run), measured 2026-11-22; repository links from the Papers with Code archive attributed under CC BY-SA, no affiliation; bodies inline only under a permissive licence as GitHub detects it. Not a reproduction of the paper's results.

"Ran" means the function executed on a synthesized input and returned a value. It is not a reproduction of the paper's results. This record, dated, as JSON: /api/ran/2312.00732.json · badge: /api/ran/2312.00732.svg

Each repository's role is a label: a classifier's output recorded on the link, or, where no role is recorded, what the Papers with Code archive marked. The line under it names its basis. Neither says who wrote the code.

RepositoryRoleRan
lkeab/gaussian-grouping canonical (classifier label)
classifier basis: census
9 ran · 4 no recorded run
FunctionStatusWhere it lives
crop_using_bbox Ran
ran; the contract check could not run (RAISES, DEP_MISSING or TIMEOUT)
lkeab/gaussian-grouping/edit_object_inpaint.py
code served (permissive licence) · get_code("15e3a8865c618243")
How it ran

Ran in Syntology's sandbox on this synthesized input (drafted by a language model, Claude Haiku 4.5, prompt python_fixture_v1, through the batch API) and returned Tensor, ndim 3, dtype family int, shape (2, 3, 4) as a later run recorded it. The image it ran in was not recorded. If your copy returns the same type, ndim and shape on these inputs, it matched what we observed; nothing here checks that. It does NOT mean the output is correct: nothing has compared it to the paper.

import torch

torch.manual_seed(0)

# Create a small image tensor
# Shape: (batch, height, width) = (2, 6, 8)
image = torch.randint(0, 256, (2, 6, 8), dtype=torch.uint8)

# Define a bounding box: (xmin, ymin, xmax, ymax)
# Using coordinates that fit within the image dimensions
bbox = (1, 1, 4, 3)

ARGS = {
    "image": image,
    "bbox": bbox,
}
Setup block drafted by
a language model, Claude Haiku 4.5, prompt python_fixture_v1, through the batch API
Imports
torch — import names as written, not package names to install; a name the repository itself provides is not a package
Returned
type Tensor, ndim 3, dtype family int, shape (2, 3, 4) — shape (2, 3, 4) as a later run recorded it
dtype family
int means an integer dtype; float means any other dtype the instrument saw, bool and complex included (fixture_check.py records only these two families)
Environment
image not recorded; re-verify against syntology-harvest-sandbox:v5-deps; not recorded for this sample: the harness installs a cuda-to-cpu shim (cuda_shim.py) before every sample and no run records whether a sample needed it; seed: none by the harness; any seed is in the setup block
divide_into_patches Ran
ran; the contract check could not run (RAISES, DEP_MISSING or TIMEOUT)
lkeab/gaussian-grouping/edit_object_inpaint.py
code served (permissive licence) · get_code("baf98c893b806f3f")
How it ran

Ran in Syntology's sandbox on this synthesized input (drafted by a language model, Claude Haiku 4.5, prompt python_fixture_v1, through the batch API) and returned Tensor, ndim 5, dtype family float, shape (2, 4, 4, 4, 4) as a later run recorded it. The image it ran in was not recorded. If your copy returns the same type, ndim and shape on these inputs, it matched what we observed; nothing here checks that. It does NOT mean the output is correct: nothing has compared it to the paper.

import torch

torch.manual_seed(0)

# Create a small image tensor: (batch=2, channels=4, height=8, width=8)
image = torch.randn(2, 4, 8, 8)

# K is the number of patches per dimension
K = 2

ARGS = {
    "image": image,
    "K": K,
}
Setup block drafted by
a language model, Claude Haiku 4.5, prompt python_fixture_v1, through the batch API
Imports
torch — import names as written, not package names to install; a name the repository itself provides is not a package
Returned
type Tensor, ndim 5, dtype family float, shape (2, 4, 4, 4, 4) — shape (2, 4, 4, 4, 4) as a later run recorded it
dtype family
int means an integer dtype; float means any other dtype the instrument saw, bool and complex included (fixture_check.py records only these two families)
Environment
image not recorded; re-verify against syntology-harvest-sandbox:v5-deps; not recorded for this sample: the harness installs a cuda-to-cpu shim (cuda_shim.py) before every sample and no run records whether a sample needed it; seed: none by the harness; any seed is in the setup block
id2rgb Ran
ran, honoured the contract we drafted
lkeab/gaussian-grouping/render.py
pointer only (licence: Apache-2.0); the stored copy get_code serves, from Runsong123/COS3D, is pointer only (licence: NOASSERTION) · get_code("9acdd81f4e30bd46")
How it ran

The body of this sample is withheld: this paper's link allows the body inline (Apache-2.0), but get_code serves the stored copy, first harvested from Runsong123/COS3D, and GitHub could not classify the licence of that repository (NOASSERTION). The setup block that called it is withheld with it, because it is written against that body. source_path and origin say where the code lives, to read under the repository's own terms. A run is recorded.

l1_loss Ran
ran; our contract draft was wrong, not the code
lkeab/gaussian-grouping/utils/loss_utils.py
code served (permissive licence) · get_code("ac0e42d6fbcfbbe6")
How it ran

Ran in Syntology's sandbox on this synthesized input (drafted by a language model, Claude Haiku 4.5, prompt python_fixture_v1, through the batch API) and returned Tensor, ndim 0, dtype family float, shape () as a later run recorded it. The image it ran in was not recorded. If your copy returns the same type, ndim and shape on these inputs, it matched what we observed; nothing here checks that. It does NOT mean the output is correct: nothing has compared it to the paper.

import torch

torch.manual_seed(0)

# Create small tensors for network output and ground truth
# Both should have the same shape for the subtraction and mean operations
network_output = torch.randn(2, 4)  # batch size 2, 4 dimensions
gt = torch.randn(2, 4)  # batch size 2, 4 dimensions

ARGS = {
    "network_output": network_output,
    "gt": gt,
}
Setup block drafted by
a language model, Claude Haiku 4.5, prompt python_fixture_v1, through the batch API
Imports
torch — import names as written, not package names to install; a name the repository itself provides is not a package
Returned
type Tensor, ndim 0, dtype family float, shape () — shape () as a later run recorded it
dtype family
int means an integer dtype; float means any other dtype the instrument saw, bool and complex included (fixture_check.py records only these two families)
Environment
image not recorded; re-verify against syntology-harvest-sandbox:v5-deps; not recorded for this sample: the harness installs a cuda-to-cpu shim (cuda_shim.py) before every sample and no run records whether a sample needed it; seed: none by the harness; any seed is in the setup block
mask_to_bbox Ran
ran; our contract draft was wrong, not the code
lkeab/gaussian-grouping/edit_object_inpaint.py
code served (permissive licence) · get_code("360f2995f7789c80")
How it ran

Ran in Syntology's sandbox on this synthesized input (drafted by a language model, Claude Haiku 4.5, prompt python_fixture_v1, through the batch API) and returned tuple, ndim 1, dtype family int; no output shape is stated. The image it ran in was not recorded. If your copy returns the same type and ndim on these inputs, it matched what we observed; nothing here checks that. It does NOT mean the output is correct: nothing has compared it to the paper.

import torch

torch.manual_seed(0)

# Create a small 2D binary mask tensor
# The function expects a 2D tensor where it checks for non-zero values along dimensions
# We'll create a mask with some True values to ensure the function can find valid bounding box coordinates
mask = torch.tensor([
    [0, 0, 1, 0],
    [0, 1, 1, 0],
    [0, 0, 0, 0],
    [0, 0, 1, 1]
], dtype=torch.uint8)

ARGS = {
    "mask": mask
}
Setup block drafted by
a language model, Claude Haiku 4.5, prompt python_fixture_v1, through the batch API
Imports
torch — import names as written, not package names to install; a name the repository itself provides is not a package
Returned
type tuple, ndim 1, dtype family int — a tuple output has no array shape of its own; none is stated
dtype family
int means an integer dtype; float means any other dtype the instrument saw, bool and complex included (fixture_check.py records only these two families)
Environment
image not recorded; re-verify against syntology-harvest-sandbox:v5-deps; not recorded for this sample: the harness installs a cuda-to-cpu shim (cuda_shim.py) before every sample and no run records whether a sample needed it; seed: none by the harness; any seed is in the setup block
masked_l1_loss Ran
ran; our contract draft was wrong, not the code
lkeab/gaussian-grouping/utils/loss_utils.py
code served (permissive licence) · get_code("0ac7968bd7da45d3")
How it ran

Ran in Syntology's sandbox on this synthesized input (drafted by a language model, Claude Haiku 4.5, prompt python_fixture_v1, through the batch API) and returned Tensor, ndim 0, dtype family float, shape () as a later run recorded it. The image it ran in was not recorded. If your copy returns the same type, ndim and shape on these inputs, it matched what we observed; nothing here checks that. It does NOT mean the output is correct: nothing has compared it to the paper.

import torch

torch.manual_seed(0)

# Create small tensors with distinct dimensions
batch_size = 2
height = 4
width = 3

# network_output: model predictions, shape (batch_size, height, width)
network_output = torch.randn(batch_size, height, width)

# gt: ground truth, same shape as network_output
gt = torch.randn(batch_size, height, width)

# mask: binary mask, shape (height, width) - will be expanded in function
mask = torch.randint(0, 2, (height, width)).float()

ARGS = {
    "network_output": network_output,
    "gt": gt,
    "mask": mask,
}
Setup block drafted by
a language model, Claude Haiku 4.5, prompt python_fixture_v1, through the batch API
Imports
torch — import names as written, not package names to install; a name the repository itself provides is not a package
Returned
type Tensor, ndim 0, dtype family float, shape () — shape () as a later run recorded it
dtype family
int means an integer dtype; float means any other dtype the instrument saw, bool and complex included (fixture_check.py records only these two families)
Environment
image not recorded; re-verify against syntology-harvest-sandbox:v5-deps; not recorded for this sample: the harness installs a cuda-to-cpu shim (cuda_shim.py) before every sample and no run records whether a sample needed it; seed: none by the harness; any seed is in the setup block
readImages Ran
ran; the contract check could not run (RAISES, DEP_MISSING or TIMEOUT)
lkeab/gaussian-grouping/metrics.py
code served (permissive licence) · get_code("cdd00787894554b5")
How it ran

Ran in Syntology's sandbox on this synthesized input (drafted by a language model, Claude Haiku 4.5, prompt python_fixture_v1, through the batch API) and returned tuple. The image it ran in was not recorded. The code mentions cuda, and the harness aliases cuda to the CPU for every sample; whether this run needed that is not recorded. If your copy returns the same type on these inputs, it matched what we observed; nothing here checks that. It does NOT mean the output is correct: nothing has compared it to the paper.

import os
import tempfile
from pathlib import Path
from PIL import Image
import torch
import numpy as np

torch.manual_seed(0)

# Create temporary directories with dummy image files
temp_renders_dir = tempfile.mkdtemp()
temp_gt_dir = tempfile.mkdtemp()

renders_dir = Path(temp_renders_dir)
gt_dir = Path(temp_gt_dir)

# Create 2 small dummy images (4x3 RGB)
for i in range(2):
    # Create render image
    render_array = np.random.randint(0, 256, (3, 4, 3), dtype=np.uint8)
    render_img = Image.fromarray(render_array.transpose(1, 2, 0), mode='RGB')
    render_img.save(renders_dir / f"render_{i}.png")
    
    # Create ground truth image
    gt_array = np.random.randint(0, 256, (3, 4, 3), dtype=np.uint8)
    gt_img = Image.fromarray(gt_array.transpose(1, 2, 0), mode='RGB')
    gt_img.save(gt_dir / f"render_{i}.png")

ARGS = {
    "renders_dir": renders_dir,
    "gt_dir": gt_dir,
}
Setup block drafted by
a language model, Claude Haiku 4.5, prompt python_fixture_v1, through the batch API
Imports
PIL, numpy, torch, torchvision — import names as written, not package names to install; a name the repository itself provides is not a package
Returned
type tuple — a tuple output has no array shape of its own; none is stated
Environment
image not recorded; re-verify against syntology-harvest-sandbox:v5-deps; not recorded for this sample: the harness installs a cuda-to-cpu shim (cuda_shim.py) before every sample and no run records whether a sample needed it; seed: none by the harness; any seed is in the setup block
show_mask Ran
ran; the contract check could not run (RAISES, DEP_MISSING or TIMEOUT)
lkeab/gaussian-grouping/ext/grounded_sam.py
code served (permissive licence) · get_code("eb026ce356b1b155")
How it ran

Ran in Syntology's sandbox on this synthesized input (drafted by a language model, Claude Haiku 4.5, prompt python_fixture_v1, through the batch API) and returned ndarray, ndim 3, dtype family int, shape (4, 6, 4) as a later run recorded it. The image it ran in was not recorded. If your copy returns the same type, ndim and shape on these inputs, it matched what we observed; nothing here checks that. It does NOT mean the output is correct: nothing has compared it to the paper.

import numpy as np
from PIL import Image

# Create a small binary mask with shape (H, W)
mask = np.random.randint(0, 2, size=(4, 6)).astype(np.uint8)

# Create a small RGB image with shape (H, W, 3)
# Image should be uint8 in range [0, 255]
image = np.random.randint(0, 256, size=(4, 6, 3), dtype=np.uint8)

ARGS = {
    "mask": mask,
    "image": image,
    "random_color": True
}
Setup block drafted by
a language model, Claude Haiku 4.5, prompt python_fixture_v1, through the batch API
Imports
PIL, numpy — import names as written, not package names to install; a name the repository itself provides is not a package
Returned
type ndarray, ndim 3, dtype family int, shape (4, 6, 4) — shape (4, 6, 4) as a later run recorded it
dtype family
int means an integer dtype; float means any other dtype the instrument saw, bool and complex included (fixture_check.py records only these two families)
Environment
image not recorded; re-verify against syntology-harvest-sandbox:v5-deps; not recorded for this sample: the harness installs a cuda-to-cpu shim (cuda_shim.py) before every sample and no run records whether a sample needed it; seed: none by the harness; any seed is in the setup block
weighted_l1_loss Ran
ran; our contract draft was wrong, not the code
lkeab/gaussian-grouping/utils/loss_utils.py
code served (permissive licence) · get_code("e14045a7568f1af5")
How it ran

Ran in Syntology's sandbox on this synthesized input (drafted by a language model, Claude Haiku 4.5, prompt python_fixture_v1, through the batch API) and returned Tensor, ndim 0, dtype family float, shape () as a later run recorded it. The image it ran in was not recorded. If your copy returns the same type, ndim and shape on these inputs, it matched what we observed; nothing here checks that. It does NOT mean the output is correct: nothing has compared it to the paper.

import torch

torch.manual_seed(0)

# The function takes three tensor parameters and performs element-wise operations
# network_output: model predictions
# gt: ground truth values
# weight: per-element weights

network_output = torch.randn(2, 4)
gt = torch.randn(2, 4)
weight = torch.rand(2, 4)

ARGS = {
    "network_output": network_output,
    "gt": gt,
    "weight": weight,
}
Setup block drafted by
a language model, Claude Haiku 4.5, prompt python_fixture_v1, through the batch API
Imports
torch — import names as written, not package names to install; a name the repository itself provides is not a package
Returned
type Tensor, ndim 0, dtype family float, shape () — shape () as a later run recorded it
dtype family
int means an integer dtype; float means any other dtype the instrument saw, bool and complex included (fixture_check.py records only these two families)
Environment
image not recorded; re-verify against syntology-harvest-sandbox:v5-deps; not recorded for this sample: the harness installs a cuda-to-cpu shim (cuda_shim.py) before every sample and no run records whether a sample needed it; seed: none by the harness; any seed is in the setup block
feature_to_rgb Not yet run
unverified (harvested, no recorded run)
lkeab/gaussian-grouping/render.py
code served (permissive licence) · get_code("955c607d3685039a")
How it ran

no invocation recorded.

points_inside_convex_hull Not yet run
unverified (harvested, no recorded run)
lkeab/gaussian-grouping/edit_object_removal.py
code served (permissive licence) · get_code("edaa321e0cd0c71a")
How it ran

no invocation recorded.

removal_setup Not yet run
unverified (harvested, no recorded run)
lkeab/gaussian-grouping/edit_object_removal.py
code served (permissive licence) · get_code("b2b9bd1de73a976f")
How it ran

no invocation recorded.

visualize_obj Not yet run
unverified (harvested, no recorded run)
lkeab/gaussian-grouping/render.py
pointer only (licence: Apache-2.0); the stored copy get_code serves, from Runsong123/COS3D, is pointer only (licence: NOASSERTION) · get_code("74195777dbe870e4")
How it ran

The body of this sample is withheld: this paper's link allows the body inline (Apache-2.0), but get_code serves the stored copy, first harvested from Runsong123/COS3D, and GitHub could not classify the licence of that repository (NOASSERTION). The setup block that called it is withheld with it, because it is written against that body. source_path and origin say where the code lives, to read under the repository's own terms. No invocation recorded.

Also present in other papers' code

Of 13 samples lifted from this paper's linked repositories, 13 carry a function identity and 0 do not. Of the 13 distinct functions with an identity, 8 are also present in other papers whose code we hold (2 with at least one paper dated earlier, 2 the same month, 8 later) and 5 are only here among the 52,977 papers whose code Syntology holds (as of 2026-09-26).

Only here means: among the 52,977 papers whose code Syntology holds (51,887 with lifted samples, 1,090 through their assemblies only), as of 2026-09-26, no other paper's lifted samples or current assembly bricks contain an identical function body after normalisation.

FunctionAlso present inEarliest-datedOthers named
l1_loss 93 other papers
8 dated earlier, 9 the same month, 76 later
2308.04079 (2023-08)
3D Gaussian Splatting for Real-Time Radiance Field Rendering
2309.13101 (2023-09)
Deformable 3D Gaussians for High-Fidelity Monocular Dynamic Scene Reconstruction
2310.08528 (2023-10)
4D Gaussian Splatting for Real-Time Dynamic Scene Rendering
and 90 more not named here
readImages 54 other papers
6 dated earlier, 4 the same month, 44 later
2308.04079 (2023-08)
3D Gaussian Splatting for Real-Time Radiance Field Rendering
2309.13101 (2023-09)
Deformable 3D Gaussians for High-Fidelity Monocular Dynamic Scene Reconstruction
2310.08528 (2023-10)
4D Gaussian Splatting for Real-Time Dynamic Scene Rendering
and 51 more not named here
id2rgb 3 other papers
3 later
2412.10051 (2024-12)
TSGaussian: Semantic and Depth-Guided Target-Specific Gaussian Splatting from Sparse Views
2503.14029 (2025-03)
Rethinking End-to-End 2D to 3D Scene Segmentation in Gaussian Splatting
2510.20238 (2025-10)
COS3D: Collaborative Open-Vocabulary 3D Segmentation
visualize_obj 3 other papers
3 later
2412.10051 (2024-12)
TSGaussian: Semantic and Depth-Guided Target-Specific Gaussian Splatting from Sparse Views
2503.14029 (2025-03)
Rethinking End-to-End 2D to 3D Scene Segmentation in Gaussian Splatting
2510.20238 (2025-10)
COS3D: Collaborative Open-Vocabulary 3D Segmentation
feature_to_rgb 2 other papers
2 later
2412.10051 (2024-12)
TSGaussian: Semantic and Depth-Guided Target-Specific Gaussian Splatting from Sparse Views
2503.14029 (2025-03)
Rethinking End-to-End 2D to 3D Scene Segmentation in Gaussian Splatting
masked_l1_loss 2 other papers
2 later
2412.10051 (2024-12)
TSGaussian: Semantic and Depth-Guided Target-Specific Gaussian Splatting from Sparse Views
2503.14029 (2025-03)
Rethinking End-to-End 2D to 3D Scene Segmentation in Gaussian Splatting
show_mask 2 other papers
2 later
2412.10051 (2024-12)
TSGaussian: Semantic and Depth-Guided Target-Specific Gaussian Splatting from Sparse Views
2503.14029 (2025-03)
Rethinking End-to-End 2D to 3D Scene Segmentation in Gaussian Splatting
weighted_l1_loss 2 other papers
2 later
2412.10051 (2024-12)
TSGaussian: Semantic and Depth-Guided Target-Specific Gaussian Splatting from Sparse Views
2503.14029 (2025-03)
Rethinking End-to-End 2D to 3D Scene Segmentation in Gaussian Splatting

Presence, not direction: an identical function body in two papers' repositories does not say who wrote it, which came from which, or that either paper's method rests on it. Matching is exact after normalisation, so a renamed or edited function is not matched. Dates are when papers appeared, not when code was written. Only .py files the harvester lifted are covered; notebooks are not read. A new-style arXiv id YYMM.NNNN dates a paper to the month 20YY-MM; otherwise Paper.year dates it to the year; otherwise it is undated. Dates are when the paper appeared, not when the code was written. Read 2026-11-22.

Repositories linked to this paper

Some links come from the archived Papers with Code dataset (CC BY-SA 4.0): attribution and licence.

Abstract

The recent Gaussian Splatting achieves high-quality and real-time novel-view synthesis of the 3D scenes. However, it is solely concentrated on the appearance and geometry modeling, while lacking in fine-grained object-level scene understanding. To address this issue, we propose Gaussian Grouping, which extends Gaussian Splatting to jointly reconstruct and segment anything in open-world 3D scenes. We augment each Gaussian with a compact Identity Encoding, allowing the Gaussians to be grouped according to their object instance or stuff membership in the 3D scene. Instead of resorting to expensive 3D labels, we supervise the Identity Encodings during the differentiable rendering by leveraging the 2D mask predictions by Segment Anything Model (SAM), along with introduced 3D spatial consistency regularization. Compared to the implicit NeRF representation, we show that the discrete and grouped 3D Gaussians can reconstruct, segment and edit anything in 3D with high visual quality, fine granularity and efficiency. Based on Gaussian Grouping, we further propose a local Gaussian Editing scheme, which shows efficacy in versatile scene editing applications, including 3D object removal, inpainting, colorization, style transfer and scene recomposition. Our code and models are at https://github.com/lkeab/gaussian-grouping.

For agents

The same record, over MCP at https://syntology.ai/mcp:

get_harvested_code_for_paper("2312.00732")
get_code_for_paper("2312.00732")
have("2312.00732")

The run record, dated, one paper per request, free:

curl https://syntology.ai/api/ran/2312.00732.json

A badge for a README (the split and the date, never a ratio):

[![Syntology run record](https://syntology.ai/api/ran/2312.00732.svg)](https://syntology.ai/paper/2312.00732)

Connect an agent — have() is free.