SYNTOLOGY HomeExplorerAtlasCodeMethodologyAboutDevelopersFeedPricing
Paper · 1905.02722 · 2019

Inverse Rendering for Complex Indoor Scenes: Shape, Spatially-Varying Lighting and SVBRDF from a Single Image

arXiv · PDF · Open in the Atlas

Code that ran

7 of 7 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; 0 honoured the contract we drafted, 0 violated it and 4 have no contract verdict; 3 failures 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/1905.02722.json · badge: /api/ran/1905.02722.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
lzqsd/InverseRenderingOfIndoorScene canonical (classifier label)
classifier basis: census
7 ran · 0 no recorded run
FunctionStatusWhere it lives
BatchRankingLoss Ran
ran on a synthesized input
lzqsd/InverseRenderingOfIndoorScene/models.py
code served (permissive licence) · get_code("8ca1216758404176")
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. 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 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
import numpy as np

torch.manual_seed(0)

# Create albedoPred: batch_size=2, height=4, width=3, channels=3
# The function accesses albedoPred.size(1) and albedoPred.size(2) for height and width
albedoPred = torch.randn(2, 4, 3, 3)

# eqPoint: numpy array of shape (num_eq_pairs, 4) with values [r1, c1, r2, c2]
# These are indices into the height x width grid, so they must be < 4 and < 3
eqPoint = np.array([[0, 0, 1, 1], [2, 1, 3, 2]], dtype=np.int64)

# eqWeight: numpy array of shape (num_eq_pairs,) with float weights
eqWeight = np.array([0.5, 0.3], dtype=np.float32)

# darkerPoint: numpy array of shape (num_darker_pairs, 4) with values [r1, c1, r2, c2]
darkerPoint = np.array([[1, 0, 2, 2], [0, 2, 3, 1]], dtype=np.int64)

# darkerWeight: numpy array of shape (num_darker_pairs,) with float weights
darkerWeight = np.array([0.4, 0.6], dtype=np.float32)

ARGS = {
    'albedoPred': albedoPred,
    'eqPoint': eqPoint,
    'eqWeight': eqWeight,
    'darkerPoint': darkerPoint,
    'darkerWeight': darkerWeight,
}
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 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
LSregress Ran
ran on a synthesized input
lzqsd/InverseRenderingOfIndoorScene/models.py
code served (permissive licence) · get_code("a3d957748a35ddcc")
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)

# Create tiny tensors with batch size 2
# pred and gt need to be reshaped to (nb, -1) so they can have any shape
# origin is used in the final computation and needs to match pred's original shape
batch_size = 2
feature_dim = 4

pred = torch.randn(batch_size, feature_dim)
gt = torch.randn(batch_size, feature_dim)
origin = torch.randn(batch_size, feature_dim)

ARGS = {
    "pred": pred,
    "gt": gt,
    "origin": origin,
}
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
LSregressDiffSpec Ran
ran on a synthesized input
lzqsd/InverseRenderingOfIndoorScene/models.py
code served (permissive licence) · get_code("b9cc21870b03e16b")
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 5, 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)

# Analyze the function to understand what shapes are needed
# The function expects:
# - diff, spec, imOrig, diffOrig, specOrig to be 4D tensors (nb, nc, nh, nw)
# - Operations: element-wise multiplication, reshaping, reductions
# - imOrig is compared with 0.9, so values should be in [0, 1]

nb = 2  # batch size
nc = 3  # number of channels (RGB)
nh = 4  # height
nw = 5  # width

# Create tensors with appropriate values
# diff: diffuse component tensor
diff = torch.randn(nb, nc, nh, nw).abs()

# spec: specular component tensor
spec = torch.randn(nb, nc, nh, nw).abs()

# imOrig: original image (values should mostly be < 0.9 for meaningful mask)
imOrig = torch.rand(nb, nc, nh, nw) * 0.95

# diffOrig: original diffuse component
diffOrig = torch.rand(nb, nc, nh, nw)

# specOrig: original specular component
specOrig = torch.rand(nb, nc, nh, nw)

ARGS = {
    "diff": diff,
    "spec": spec,
    "imOrig": imOrig,
    "diffOrig": diffOrig,
    "specOrig": specOrig,
}
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 5, 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
compute_whdr Ran
ran on a synthesized input
lzqsd/InverseRenderingOfIndoorScene/CompareWHDR.py
code served (permissive licence) · get_code("8a12b5688f68f4d8")
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 numpy as np

# Create a small reflectance image (height=10, width=10, channels=3)
reflectance = np.random.rand(10, 10, 3).astype(np.float32)

# Create judgements dict with intrinsic_points and intrinsic_comparisons
judgements = {
    'intrinsic_points': [
        {
            'id': 'p1',
            'x': 0.2,  # normalized x coordinate
            'y': 0.3,  # normalized y coordinate
            'opaque': True
        },
        {
            'id': 'p2',
            'x': 0.5,
            'y': 0.6,
            'opaque': True
        },
        {
            'id': 'p3',
            'x': 0.7,
            'y': 0.8,
            'opaque': False
        }
    ],
    'intrinsic_comparisons': [
        {
            'point1': 'p1',
            'point2': 'p2',
            'darker': '1',  # point1 is darker
            'darker_score': 0.9
        },
        {
            'point1': 'p2',
            'point2': 'p3',
            'darker': 'E',  # equal
            'darker_score': 0.8
        },
        {
            'point1': 'p1',
            'point2': 'p2',
            'darker': '2',  # point2 is darker
            'darker_score': 0.7
        }
    ]
}

ARGS = {
    'reflectance': reflectance,
    'judgements': judgements,
    'delta': 0.1
}
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 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
get_valid_idx Ran
ran; the contract check could not run (RAISES, DEP_MISSING or TIMEOUT)
lzqsd/InverseRenderingOfIndoorScene/BilateralGrid.py
pointer only (licence: MIT); the stored copy get_code serves, from jiahao000/mosaicfusion, is pointer only (licence: NOASSERTION) · get_code("04c4101fccb5c55b")
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 jiahao000/mosaicfusion, 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.

rgb2yuv Ran
ran; the contract check could not run (RAISES, DEP_MISSING or TIMEOUT)
lzqsd/InverseRenderingOfIndoorScene/BilateralGrid.py
pointer only (licence: MIT); the stored copy get_code serves, from jiahao000/mosaicfusion, is pointer only (licence: NOASSERTION) · get_code("ec22a04e890fb286")
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 jiahao000/mosaicfusion, 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.

yuv2rgb Ran
ran; the contract check could not run (RAISES, DEP_MISSING or TIMEOUT)
lzqsd/InverseRenderingOfIndoorScene/BilateralGrid.py
pointer only (licence: MIT); the stored copy get_code serves, from jiahao000/mosaicfusion, is pointer only (licence: NOASSERTION) · get_code("5d9850cf710de517")
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 jiahao000/mosaicfusion, 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.

Also present in other papers' code

Of 7 samples lifted from this paper's linked repositories, 7 carry a function identity and 0 do not. Of the 7 distinct functions with an identity, 3 are also present in other papers whose code we hold (0 with at least one paper dated earlier, 0 the same month, 3 later) and 4 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_valid_idx 4 other papers
4 later
2212.05853 (2022-12)
DeepCut: Unsupervised Segmentation using Graph Neural Networks Clustering
2212.07834 (2022-12)
Unsupervised Object Localization: Observing the Background to Discover Objects
2309.13042 (2023-09)
MosaicFusion: Diffusion Models as Data Augmenters for Large Vocabulary Instance Segmentation
and 1 more not named here
rgb2yuv 3 other papers
3 later
2212.05853 (2022-12)
DeepCut: Unsupervised Segmentation using Graph Neural Networks Clustering
2212.07834 (2022-12)
Unsupervised Object Localization: Observing the Background to Discover Objects
2309.13042 (2023-09)
MosaicFusion: Diffusion Models as Data Augmenters for Large Vocabulary Instance Segmentation
yuv2rgb 3 other papers
3 later
2212.05853 (2022-12)
DeepCut: Unsupervised Segmentation using Graph Neural Networks Clustering
2212.07834 (2022-12)
Unsupervised Object Localization: Observing the Background to Discover Objects
2309.13042 (2023-09)
MosaicFusion: Diffusion Models as Data Augmenters for Large Vocabulary Instance Segmentation

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 propose a deep inverse rendering framework for indoor scenes. From a single RGB image of an arbitrary indoor scene, we create a complete scene reconstruction, estimating shape, spatially-varying lighting, and spatially-varying, non-Lambertian surface reflectance. To train this network, we augment the SUNCG indoor scene dataset with real-world materials and render them with a fast, high-quality, physically-based GPU renderer to create a large-scale, photorealistic indoor dataset. Our inverse rendering network incorporates physical insights -- including a spatially-varying spherical Gaussian lighting representation, a differentiable rendering layer to model scene appearance, a cascade structure to iteratively refine the predictions and a bilateral solver for refinement -- allowing us to jointly reason about shape, lighting, and reflectance. Experiments show that our framework outperforms previous methods for estimating individual scene components, which also enables various novel applications for augmented reality, such as photorealistic object insertion and material editing. Code and data will be made publicly available.

For agents

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

get_harvested_code_for_paper("1905.02722")
get_code_for_paper("1905.02722")
have("1905.02722")

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

curl https://syntology.ai/api/ran/1905.02722.json

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

[![Syntology run record](https://syntology.ai/api/ran/1905.02722.svg)](https://syntology.ai/paper/1905.02722)

Connect an agent — have() is free.