SYNTOLOGY HomeExplorerAtlasCodeMethodologyAboutDevelopersFeedPricing
Paper · 2003.10432 · 2020

Atlas: End-to-End 3D Scene Reconstruction from Posed Images

arXiv · PDF · Open in the Atlas

Code that ran

11 of 12 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 7 have no contract verdict; 3 failures of our instrument; 1 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/2003.10432.json · badge: /api/ran/2003.10432.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
magicleap/Atlas community (classifier label)
classifier basis not recorded on the link
11 ran · 1 no recorded run
FunctionStatusWhere it lives
conv1x1x1 Ran
ran; our contract draft was wrong, not the code
magicleap/Atlas/atlas/backbone3d.py
code served (permissive licence) · get_code("871844ec996c4499")
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 Conv3d. 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 torch
import torch.nn as nn

torch.manual_seed(0)

ARGS = {
    'in_planes': 3,
    'out_planes': 64,
    'stride': 1
}
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 Conv3d — a Conv3d output has no array shape of its own; none is stated — the instrument coerces any output to an array; for a Conv3d output the recorded ndim 0 and dtype family float describe that coercion, not the output, and are not 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
conv3x3x3 Ran
ran on a synthesized input
magicleap/Atlas/atlas/backbone3d.py
code served (permissive licence) · get_code("345667d343bb0d99")
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 Conv3d. 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 torch
import torch.nn as nn

torch.manual_seed(0)

ARGS = {
    "in_planes": 3,
    "out_planes": 64,
    "stride": 1,
    "groups": 1,
    "dilation": 1,
}
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 Conv3d — a Conv3d output has no array shape of its own; none is stated — the instrument coerces any output to an array; for a Conv3d output the recorded ndim 0 and dtype family float describe that coercion, not the output, and are not 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
coordinates Ran
ran on a synthesized input
magicleap/Atlas/atlas/tsdf.py
code served (permissive licence) · get_code("880183f37459c9c0")
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 2, dtype family int, shape (3, 24) as a later run recorded it. 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, 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 voxel_dim (tuple of 3 ints) and device
# It creates a 3D meshgrid and returns stacked coordinates
# voxel_dim should be a tuple of 3 integers specifying grid dimensions

voxel_dim = (2, 3, 4)  # Small distinct sizes for nx, ny, nz
device = torch.device('cpu')

ARGS = {
    'voxel_dim': voxel_dim,
    'device': device
}
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 2, dtype family int, shape (3, 24) — shape (3, 24) 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
depth_to_world Ran
ran on a synthesized input
magicleap/Atlas/atlas/tsdf.py
code served (permissive licence) · get_code("d673fe1180f63a4c")
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 float, shape (3, 2, 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 torch

torch.manual_seed(0)

# Create a 3x4 projection matrix
projection = torch.randn(3, 4, dtype=torch.float32)

# Create a small depth map (hxw)
depth = torch.randn(2, 3, dtype=torch.float32) + 1.0  # Add 1.0 to avoid division by zero issues

ARGS = {
    "projection": projection,
    "depth": depth,
}
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 float, shape (3, 2, 3) — shape (3, 2, 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
get_norm_3d Ran
ran on a synthesized input
magicleap/Atlas/atlas/backbone3d.py
code served (permissive licence) · get_code("10e977bbee5162f4")
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 BatchNorm3d. 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 torch
import torch.nn as nn

torch.manual_seed(0)

# Test case 1: norm as string 'BN'
ARGS = {
    'norm': 'BN',
    'out_channels': 64
}
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 BatchNorm3d — a BatchNorm3d output has no array shape of its own; none is stated — the instrument coerces any output to an array; for a BatchNorm3d output the recorded ndim 0 and dtype family float describe that coercion, not the output, and are not 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
l1 Ran
ran, honoured the contract we drafted
magicleap/Atlas/atlas/evaluation.py
code served (permissive licence) · get_code("60530e4179d62619")
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 float, 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
import torch.nn as nn

torch.manual_seed(0)

class TSDF:
    def __init__(self, voxel_size, origin, tsdf_vol):
        self.voxel_size = voxel_size
        self.origin = origin
        self.tsdf_vol = tsdf_vol

# Create prediction TSDF
voxel_size_pred = 0.01
origin_pred = torch.tensor([0.0, 0.0, 0.0])
tsdf_vol_pred = torch.randn(4, 5, 6)

tsdf_pred = TSDF(voxel_size_pred, origin_pred, tsdf_vol_pred)

# Create target TSDF with same properties
voxel_size_trgt = 0.01
origin_trgt = torch.tensor([0.0, 0.0, 0.0])
# Set some values to be >= 1 to test the mask
tsdf_vol_trgt = torch.randn(4, 5, 6)
tsdf_vol_trgt[0, 0, 0] = 1.5  # This should be masked out

tsdf_trgt = TSDF(voxel_size_trgt, origin_trgt, tsdf_vol_trgt)

ARGS = {
    "tsdf_pred": tsdf_pred,
    "tsdf_trgt": tsdf_trgt
}
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 float, 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
load_info_json Ran
ran on a synthesized input
magicleap/Atlas/atlas/data.py
code served (permissive licence) · get_code("e0258b4df51c5013")
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 dict. 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 json
import tempfile
import os

# Create a temporary JSON file with sample data
temp_dir = tempfile.mkdtemp()
json_file = os.path.join(temp_dir, "info.json")

# Test case 1: info with instances as a dict
info_data = {
    "instances": {"0": {"id": 0, "name": "obj0"}, "1": {"id": 1, "name": "obj1"}},
    "metadata": "some metadata"
}

with open(json_file, 'w') as f:
    json.dump(info_data, f)

ARGS = {
    "json_file": json_file
}
Setup block drafted by
a language model, Claude Haiku 4.5, prompt python_fixture_v1, through the batch API
Imports
none outside the standard library — import names as written, not package names to install; a name the repository itself provides is not a package
Returned
type dict — a dict output has no array shape of its own; none is stated — the instrument coerces any output to an array; for a dict output the recorded ndim 0 and dtype family float describe that coercion, not the output, and are not 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
load_scannet_label_mapping Ran
ran on a synthesized input
magicleap/Atlas/atlas/datasets/scannet.py
code served (permissive licence) · get_code("963283e33428e65d")
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 dict. 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 csv
import os
import tempfile

# Create a temporary directory and a mock scannetv2-labels.combined.tsv file
temp_dir = tempfile.mkdtemp()

# Create the TSV file with header and sample data
tsv_path = os.path.join(temp_dir, 'scannetv2-labels.combined.tsv')
with open(tsv_path, 'w', newline='') as f:
    writer = csv.writer(f, delimiter='\t')
    writer.writerow(['id', 'name'])  # header
    writer.writerow(['1', 'wall'])
    writer.writerow(['2', 'chair'])
    writer.writerow(['22', 'books'])
    writer.writerow(['5', 'floor'])
    writer.writerow(['7', 'door'])

ARGS = {
    'path': temp_dir
}
Setup block drafted by
a language model, Claude Haiku 4.5, prompt python_fixture_v1, through the batch API
Imports
none outside the standard library — import names as written, not package names to install; a name the repository itself provides is not a package
Returned
type dict — a dict output has no array shape of its own; none is stated — the instrument coerces any output to an array; for a dict output the recorded ndim 0 and dtype family float describe that coercion, not the output, and are not 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
log_transform Ran
ran; our contract draft was wrong, not the code
magicleap/Atlas/atlas/heads3d.py
code served (permissive licence) · get_code("cc965a768c8eb8cb")
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 2, dtype family float, shape (2, 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)

# The function takes x (a tensor with sign() and abs() methods)
# and shift (a numeric parameter for scaling)
# It performs: x.sign() * (1 + x.abs() / shift).log()
# So x needs to be a tensor with these methods available

x = torch.randn(2, 4)
shift = 1

ARGS = {
    "x": x,
    "shift": shift,
}
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 2, dtype family float, shape (2, 4) — shape (2, 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
map_frame Ran
ran on a synthesized input
magicleap/Atlas/atlas/data.py
code served (permissive licence) · get_code("cd19ec43da6ecc6e")
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 dict. 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
from PIL import Image
import tempfile
import os

# Create temporary directory for test images
temp_dir = tempfile.mkdtemp()

# Create a simple test image (RGB)
test_image = Image.new('RGB', (64, 48), color=(100, 150, 200))
image_path = os.path.join(temp_dir, 'test_image.png')
test_image.save(image_path)

# Create a simple depth image (grayscale)
depth_image = Image.new('L', (64, 48), color=100)
depth_path = os.path.join(temp_dir, 'test_depth.png')
depth_image.save(depth_path)

# Create a simple instance segmentation image
instance_image = Image.new('L', (64, 48), color=50)
instance_path = os.path.join(temp_dir, 'test_instance.png')
instance_image.save(instance_path)

# Construct frame dict with all required metadata
frame = {
    'file_name_image': image_path,
    'file_name_depth': depth_path,
    'file_name_instance': instance_path,
    'intrinsics': [[500.0, 0.0, 32.0],
                   [0.0, 500.0, 24.0],
                   [0.0, 0.0, 1.0]],
    'pose': [[1.0, 0.0, 0.0, 0.0],
             [0.0, 1.0, 0.0, 0.0],
             [0.0, 0.0, 1.0, 0.0],
             [0.0, 0.0, 0.0, 1.0]],
    'frame_id': 0,
    'scene_id': 'test_scene'
}

ARGS = {
    'frame': frame,
    'frame_types': ['depth', 'semseg']
}
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 dict — a dict output has no array shape of its own; none is stated — the instrument coerces any output to an array; for a dict output the recorded ndim 0 and dtype family float describe that coercion, not the output, and are not 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
pad_scannet Ran
ran; our contract draft was wrong, not the code
magicleap/Atlas/atlas/transforms.py
pointer only (licence: Apache-2.0); the stored copy get_code serves, from sercharles/cn-rma, is pointer only (licence: NOASSERTION) · get_code("930cb92f559f5607")
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 sercharles/cn-rma, 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.

transform_space Not yet run
unverified (harvested, no recorded run)
magicleap/Atlas/atlas/transforms.py
code served (permissive licence) · get_code("41439e78eb397c28")
How it ran

no invocation recorded.

Also present in other papers' code

Of 12 samples lifted from this paper's linked repositories, 12 carry a function identity and 0 do not. Of the 12 distinct functions with an identity, 4 are also present in other papers whose code we hold (1 with at least one paper dated earlier, 0 the same month, 4 later) and 8 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
conv1x1x1 6 other papers
1 dated earlier, 5 later · 1 only as an assembly brick
1711.09577 (2017-11)
Can Spatiotemporal 3D CNNs Retrace the History of 2D CNNs and ImageNet?
2004.04968 (2020-04)
Would Mega-scale Datasets Further Enhance Spatiotemporal 3D CNNs?
2107.07695 (2021-07)
Self-supervised Representation Learning Framework for Remote Physiological Measurement Using Spatiotemporal Augmentation Loss
and 3 more not named here
log_transform 2 other papers
2 later
2112.00236 (2021-12)
VoRTX: Volumetric 3D Reconstruction With Transformers for Voxelwise View Selection and Fusion
2301.13510 (2023-01)
3D FORMER: MONOCULAR SCENE RECONSTRUCTION WITH 3D SDF TRANSFORMERS
l1 1 other paper
1 later
2112.00236 (2021-12)
VoRTX: Volumetric 3D Reconstruction With Transformers for Voxelwise View Selection and Fusion
pad_scannet 1 other paper
1 later
2403.04198 (2024-03)
This CVPR paper is the Open Access version, provided by the Computer Vision Foundation. Except for this watermark, it is identical to the accepted version; the final published version of the proceedings is available on IEEE Xplore

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

We present an end-to-end 3D reconstruction method for a scene by directly regressing a truncated signed distance function (TSDF) from a set of posed RGB images. Traditional approaches to 3D reconstruction rely on an intermediate representation of depth maps prior to estimating a full 3D model of a scene. We hypothesize that a direct regression to 3D is more effective. A 2D CNN extracts features from each image independently which are then back-projected and accumulated into a voxel volume using the camera intrinsics and extrinsics. After accumulation, a 3D CNN refines the accumulated features and predicts the TSDF values. Additionally, semantic segmentation of the 3D model is obtained without significant computation. This approach is evaluated on the Scannet dataset where we significantly outperform state-of-the-art baselines (deep multiview stereo followed by traditional TSDF fusion) both quantitatively and qualitatively. We compare our 3D semantic segmentation to prior methods that use a depth sensor since no previous work attempts the problem with only RGB input.

For agents

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

get_harvested_code_for_paper("2003.10432")
get_code_for_paper("2003.10432")
have("2003.10432")

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

curl https://syntology.ai/api/ran/2003.10432.json

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

[![Syntology run record](https://syntology.ai/api/ran/2003.10432.svg)](https://syntology.ai/paper/2003.10432)

Connect an agent — have() is free.