SYNTOLOGY HomeExplorerAtlasCodeMethodologyAboutDevelopersFeedPricing
Paper · 2303.12012 · CVPR · 2023

NeAT: Learning Neural Implicit Surfaces with Arbitrary Topologies from Multi-view Images

Weikai Chen, Xiaoxu Meng

arXiv · PDF · Open in the Atlas

Code that ran

10 of 10 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, 1 violated it and 7 have no contract verdict; 1 failure of our instrument; 0 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/2303.12012.json · badge: /api/ran/2303.12012.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
xmeng525/NeAT canonical (classifier label)
classifier basis not recorded on the link
10 ran · 0 no recorded run
FunctionStatusWhere it lives
extract_fields Ran
ran on a synthesized input
xmeng525/NeAT/models/renderer.py
code served (permissive licence) · get_code("c3422de397b2766c")
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 float, shape (128, 128, 128) 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
import numpy as np

torch.manual_seed(0)

# Define a simple query function that takes points and returns scalar values
def query_func(pts):
    """
    Takes pts of shape (N, M, 3) and returns scalar values of shape (N, M)
    This is a simple stand-in that computes the norm of each point.
    """
    # pts shape: (N, M, 3)
    # Return shape: (N, M)
    return torch.norm(pts, dim=-1)

ARGS = {
    "bound_min": np.array([-1.0, -1.0, -1.0]),
    "bound_max": np.array([1.0, 1.0, 1.0]),
    "resolution": 128,
    "query_func": query_func,
}
Setup block drafted by
a language model, Claude Haiku 4.5, prompt python_fixture_v1, through the batch API
Imports
numpy, torch — 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 float, shape (128, 128, 128) — shape (128, 128, 128) 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
generate_cam_positions Ran
ran on a synthesized input
xmeng525/NeAT/data_generation/render_pc.py
code served (permissive licence) · get_code("e1ca3ad555a6a160")
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. 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 numpy as np

ARGS = {
    'elev_start': 0.0,
    'elev_end': 90.0,
    'elev_step': 3,
    'elev_offset': 5.0,
    'azim_start': 0.0,
    'azim_end': 360.0,
    'azim_step': 4,
}
Setup block drafted by
a language model, Claude Haiku 4.5, prompt python_fixture_v1, through the batch API
Imports
numpy — 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
generate_scale_mat Ran
ran on a synthesized input
xmeng525/NeAT/data_generation/render_pc.py
code served (permissive licence) · get_code("860258054d757c13")
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 2, dtype family float, shape (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 numpy as np

def generate_scale_mat(res):
    scale_mat = np.eye(4)
    scale_mat[0, 0] *= (res - 1) / 2
    scale_mat[1, 1] *= (res - 1) / 2
    scale_mat[0, 2] += (res - 1) / 2
    scale_mat[1, 2] += (res - 1) / 2
    return scale_mat

ARGS = {
    "res": 5
}
Setup block drafted by
a language model, Claude Haiku 4.5, prompt python_fixture_v1, through the batch API
Imports
numpy — import names as written, not package names to install; a name the repository itself provides is not a package
Returned
type ndarray, ndim 2, dtype family float, shape (4, 4) — shape (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
getModelViewMatrix Ran
ran on a synthesized input
xmeng525/NeAT/data_generation/cam_utils.py
code served (permissive licence) · get_code("00797100cfd918f4")
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 2, dtype family float, shape (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 numpy as np

ARGS = {
    "eye": np.asarray([1.0, 2.0, 3.0]),
    "up": np.asarray([0.0, 1.0, 0.0])
}
Setup block drafted by
a language model, Claude Haiku 4.5, prompt python_fixture_v1, through the batch API
Imports
numpy — import names as written, not package names to install; a name the repository itself provides is not a package
Returned
type ndarray, ndim 2, dtype family float, shape (4, 4) — shape (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
getPinHoleMatrix Ran
ran on a synthesized input
xmeng525/NeAT/data_generation/cam_utils.py
code served (permissive licence) · get_code("ded05e783b37fc4d")
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 2, dtype family float, shape (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 numpy as np

ARGS = {
    "scalings": [1, 1, 1]
}
Setup block drafted by
a language model, Claude Haiku 4.5, prompt python_fixture_v1, through the batch API
Imports
numpy — import names as written, not package names to install; a name the repository itself provides is not a package
Returned
type ndarray, ndim 2, dtype family float, shape (4, 4) — shape (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
get_embedder Ran
ran; our contract draft was wrong, not the code
xmeng525/NeAT/models/embedder.py
code served (permissive licence) · get_code("bb0e12df9c0d2a05")
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 float; 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)

ARGS = {
    'multires': 4,
    'input_dims': 3
}
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 float — 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
load_K_Rt_from_P Ran
ran, honoured the contract we drafted
xmeng525/NeAT/models/dataset.py
pointer only (licence: MIT); the stored copy get_code serves, from 728388808/alpha-NeuS, is pointer only (licence: NONE) · get_code("69986a1a9ebdac78")
How it ran

The body of this sample is withheld: this paper's link allows the body inline (MIT), but get_code serves the stored copy, first harvested from 728388808/alpha-NeuS, and GitHub detected no licence in that repository when it was checked. 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.

normalize Ran
ran, violated the contract we drafted
xmeng525/NeAT/data_generation/cam_utils.py
code served (permissive licence) · get_code("a3ba2d7efec77ee5")
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 1, dtype family float, shape (3,) 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

# The function normalize takes a vector v of shape (3,) and normalizes it
# by dividing by its L2 norm. We need to provide a numpy array of shape (3,).

# Create a simple 3-element vector
v = np.array([3.0, 4.0, 0.0])

ARGS = {
    "v": v
}
Setup block drafted by
a language model, Claude Haiku 4.5, prompt python_fixture_v1, through the batch API
Imports
numpy — import names as written, not package names to install; a name the repository itself provides is not a package
Returned
type ndarray, ndim 1, dtype family float, shape (3,) — shape (3,) 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
read_ply_xyzrgb Ran
ran on a synthesized input
xmeng525/NeAT/data_generation/render_pc.py
code served (permissive licence) · get_code("418e1658c345c43a")
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 2, dtype family float, shape (3, 6) 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 os
import tempfile
import numpy as np
from plyfile import PlyData, PlyElement

# Create a temporary PLY file with valid XYZ RGB data
temp_dir = tempfile.mkdtemp()
temp_ply_path = os.path.join(temp_dir, "test.ply")

# Create sample vertex data with x, y, z, red, green, blue
num_vertices = 3
vertex_data = np.zeros(num_vertices, dtype=[('x', 'f4'), ('y', 'f4'), ('z', 'f4'),
                                             ('red', 'u1'), ('green', 'u1'), ('blue', 'u1')])
vertex_data['x'] = np.array([0.1, 0.2, 0.3], dtype=np.float32)
vertex_data['y'] = np.array([0.4, 0.5, 0.6], dtype=np.float32)
vertex_data['z'] = np.array([0.7, 0.8, 0.9], dtype=np.float32)
vertex_data['red'] = np.array([255, 128, 64], dtype=np.uint8)
vertex_data['green'] = np.array([100, 150, 200], dtype=np.uint8)
vertex_data['blue'] = np.array([50, 75, 125], dtype=np.uint8)

# Create and write PLY file
vertex_el = PlyElement.describe(vertex_data, 'vertex')
PlyData([vertex_el]).write(temp_ply_path)

ARGS = {
    "filename": temp_ply_path
}
Setup block drafted by
a language model, Claude Haiku 4.5, prompt python_fixture_v1, through the batch API
Imports
numpy, plyfile — import names as written, not package names to install; a name the repository itself provides is not a package
Returned
type ndarray, ndim 2, dtype family float, shape (3, 6) — shape (3, 6) 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
remove_nan_from_mesh Ran
ran on a synthesized input
xmeng525/NeAT/models/mesh_utils.py
code served (permissive licence) · get_code("ffb7dd9ebf9adab7")
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. 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 numpy as np

# Create sample mesh data with some NaN values
verts = np.array([
    [0.0, 0.1, 0.2],
    [1.0, 1.1, 1.2],
    [np.nan, 2.1, 2.2],  # vertex with NaN
    [3.0, 3.1, 3.2],
    [4.0, np.nan, 4.2],  # vertex with NaN
    [5.0, 5.1, 5.2],
])

faces = np.array([
    [0, 1, 2],  # contains vertex 2 (NaN)
    [1, 3, 5],  # all valid
    [3, 4, 5],  # contains vertex 4 (NaN)
    [0, 1, 3],  # all valid
])

ARGS = {
    "verts": verts,
    "faces": faces,
}
Setup block drafted by
a language model, Claude Haiku 4.5, prompt python_fixture_v1, through the batch API
Imports
numpy — 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

Also present in other papers' code

Of 10 samples lifted from this paper's linked repositories, 10 carry a function identity and 0 do not. Of the 10 distinct functions with an identity, 3 are also present in other papers whose code we hold (3 with at least one paper dated earlier, 2 the same month, 2 later) and 7 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
get_embedder 22 other papers
6 dated earlier, 2 the same month, 13 later, 1 same year, order not determinable · 3 only as an assembly brick
2106.10689 (2021-06)
NeuS: Learning Neural Implicit Surfaces by Volume Rendering for Multi-view Reconstruction
2106.12052 (2021-06)
Volume Rendering of Neural Implicit Surfaces
2205.15848 (2022-05)
Geo-Neus: Geometry-Consistent Neural Implicit Surfaces Learning for Multi-view Reconstruction
and 19 more not named here
load_K_Rt_from_P 14 other papers
4 dated earlier, 1 the same month, 8 later, 1 same year, order not determinable
2106.10689 (2021-06)
NeuS: Learning Neural Implicit Surfaces by Volume Rendering for Multi-view Reconstruction
2201.05989 (2022-01)
Instant neural graphics primitives with a multiresolution hash encoding
2205.15848 (2022-05)
Geo-Neus: Geometry-Consistent Neural Implicit Surfaces Learning for Multi-view Reconstruction
and 11 more not named here
normalize 1 other paper
1 dated earlier
2010.01114 (2020-10)
Goal-GAN: Multimodal Trajectory Prediction Based on Goal Position Estimation

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

Figure 1. We show three groups of surface reconstruction from multi-view images. The front and back faces are rendered in blue and yellow respectively. Our method (left) is able to reconstruct high-fidelity and intricate surfaces of arbitrary topologies, including those non-watertight structures, e.g. the thin single-layer shoulder strap of the top (middle). In comparison, the state-of-the-art NeuS [48] method (right) can only generate watertight surfaces, resulting in thick, double-layer geometries.

For agents

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

get_harvested_code_for_paper("2303.12012")
get_code_for_paper("2303.12012")
have("2303.12012")

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

curl https://syntology.ai/api/ran/2303.12012.json

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

[![Syntology run record](https://syntology.ai/api/ran/2303.12012.svg)](https://syntology.ai/paper/2303.12012)

Connect an agent — have() is free.