SYNTOLOGY HomeExplorerAtlasCodeMethodologyAboutDevelopersFeedPricing
Paper · 1901.06447 · 2019

Learning single-image 3D reconstruction by generative modelling of shape, pose and shading

arXiv · PDF · Open in the Atlas

Code that ran

8 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; 0 honoured the contract we drafted, 0 violated it and 8 have no contract verdict; 0 failures of our instrument; 5 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/1901.06447.json · badge: /api/ran/1901.06447.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
pmh47/dirt canonical (classifier label)
classifier basis: census
8 ran · 5 no recorded run
FunctionStatusWhere it lives
rodrigues Ran
ran on a synthesized input
pmh47/dirt/dirt/matrices.py
code served (permissive licence) · get_code("f3b24a617870bcaa")
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 EagerTensor, ndim 3, dtype family float, shape (2, 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 tensorflow as tf
import numpy as np

# Set seed for reproducibility
tf.random.set_seed(0)
np.random.seed(0)

# Create a small batch of angle-axis rotation vectors
# Shape: [batch_size, 3] where batch_size=2
batch_size = 2
vectors = tf.constant([
    [0.1, 0.2, 0.3],
    [0.4, 0.5, 0.6]
], dtype=tf.float32)

ARGS = {
    'vectors': vectors,
    'name': None,
    'three_by_three': False
}
Setup block drafted by
a language model, Claude Haiku 4.5, prompt python_fixture_v1, through the batch API
Imports
numpy, tensorflow — import names as written, not package names to install; a name the repository itself provides is not a package
Returned
type EagerTensor, ndim 3, dtype family float, shape (2, 4, 4) — shape (2, 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
sample_texture Ran
ran on a synthesized input
pmh47/dirt/samples/textured.py
code served (permissive licence) · get_code("9c52a472ec1babee")
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 SymbolicTensor. 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 tensorflow as tf
import numpy as np

# Create a small texture tensor (height=4, width=4, channels=3)
texture = tf.constant(
    np.random.rand(4, 4, 3).astype(np.float32),
    dtype=tf.float32
)

# Create indices for sampling (batch_size=2, 2D coordinates)
# Indices should be in range [0, 3] for valid texture coordinates
indices = tf.constant(
    [[1.5, 1.5], [2.3, 0.7]],
    dtype=tf.float32
)

ARGS = {
    'texture': texture,
    'indices': indices,
    'mode': 'bilinear'
}
Setup block drafted by
a language model, Claude Haiku 4.5, prompt python_fixture_v1, through the batch API
Imports
numpy, tensorflow — import names as written, not package names to install; a name the repository itself provides is not a package
Returned
type SymbolicTensor — a SymbolicTensor 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
scale Ran
ran on a synthesized input
pmh47/dirt/dirt/matrices.py
code served (permissive licence) · get_code("fb7e722cbd3a246a")
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 EagerTensor, ndim 3, dtype family float, shape (2, 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 tensorflow as tf
import numpy as np

# Create a small batch of 3D scale factors
# Shape: [2, 3] - batch size 2, 3 scale factors per item
scale_factors = tf.constant([
    [1.0, 2.0, 3.0],
    [0.5, 1.5, 2.5]
], dtype=tf.float32)

ARGS = {
    'x': scale_factors,
    'name': None
}
Setup block drafted by
a language model, Claude Haiku 4.5, prompt python_fixture_v1, through the batch API
Imports
numpy, tensorflow — import names as written, not package names to install; a name the repository itself provides is not a package
Returned
type EagerTensor, ndim 3, dtype family float, shape (2, 4, 4) — shape (2, 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
split_vertices_by_face Ran
ran on a synthesized input
pmh47/dirt/dirt/lighting.py
code served (permissive licence) · get_code("8be846be750b646f")
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 tensorflow as tf
import numpy as np

# Create small sample tensors
# vertices: shape [batch, vertex_count, 3] or [batch, vertex_count, 4]
# Using shape [2, 4, 3] - batch of 2, 4 vertices, 3D coordinates
vertices_data = np.array([
    [[0.0, 0.0, 0.0],
     [1.0, 0.0, 0.0],
     [0.0, 1.0, 0.0],
     [1.0, 1.0, 0.0]],
    [[0.0, 0.0, 1.0],
     [1.0, 0.0, 1.0],
     [0.0, 1.0, 1.0],
     [1.0, 1.0, 1.0]]
], dtype=np.float32)

vertices = tf.constant(vertices_data)

# faces: shape [face_count, 3]
# Each row is indices into vertices (0-indexed)
# Using 2 triangles
faces_data = np.array([
    [0, 1, 2],
    [1, 3, 2]
], dtype=np.int32)

faces = tf.constant(faces_data)

ARGS = {
    'vertices': vertices,
    'faces': faces,
    'name': None
}
Setup block drafted by
a language model, Claude Haiku 4.5, prompt python_fixture_v1, through the batch API
Imports
numpy, tensorflow — 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
translation Ran
ran on a synthesized input
pmh47/dirt/dirt/matrices.py
code served (permissive licence) · get_code("bacb86df9c02264f")
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 EagerTensor, ndim 3, dtype family float, shape (2, 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 tensorflow as tf
import numpy as np

tf.random.set_seed(0)

# Create a batch of 3D displacement vectors
# Shape: [2, 3] - batch of 2, with 3 displacement components
x_data = tf.constant([
    [1.0, 2.0, 3.0],
    [4.0, 5.0, 6.0]
], dtype=tf.float32)

ARGS = {
    'x': x_data,
    'name': None
}
Setup block drafted by
a language model, Claude Haiku 4.5, prompt python_fixture_v1, through the batch API
Imports
numpy, tensorflow — import names as written, not package names to install; a name the repository itself provides is not a package
Returned
type EagerTensor, ndim 3, dtype family float, shape (2, 4, 4) — shape (2, 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
unit Ran
ran on a synthesized input
pmh47/dirt/samples/simple.py
code served (permissive licence) · get_code("14ff1095762d2595")
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 EagerTensor, ndim 1, dtype family float, shape (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 tensorflow as tf
import numpy as np

# Construct a valid argument for the unit function
# The function expects a vector-like object that can be converted to a tensor
# and supports the norm operation

# Create a small vector as a list
vector = [1.0, 2.0, 3.0, 4.0]

ARGS = {
    "vector": vector
}
Setup block drafted by
a language model, Claude Haiku 4.5, prompt python_fixture_v1, through the batch API
Imports
numpy, tensorflow — import names as written, not package names to install; a name the repository itself provides is not a package
Returned
type EagerTensor, ndim 1, dtype family float, shape (4,) — shape (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
unproject_pixels_to_rays Ran
ran on a synthesized input
pmh47/dirt/dirt/projection.py
code served (permissive licence) · get_code("3067554ee464d491")
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 tensorflow as tf
import numpy as np

# Set random seed for reproducibility
tf.random.set_seed(0)

# Create small test tensors
batch_size = 2
num_pixels = 3  # B1 dimension for pixel locations
image_width = 640
image_height = 480

# pixel_locations: shape [A1, ..., An, B1, ..., Bm, 2]
# Let's use shape [2, 3, 2] meaning batch=2, num_pixels=3, xy_coords=2
pixel_locations = tf.constant([
    [[100.0, 150.0], [200.0, 250.0], [300.0, 350.0]],
    [[50.0, 100.0], [150.0, 200.0], [250.0, 300.0]]
], dtype=tf.float32)

# clip_to_world_matrix: shape [A1, ..., An, 4, 4]
# Standard identity-like matrix for unprojection
# Shape [2, 4, 4] to match batch dimension of pixel_locations
clip_to_world_matrix = tf.constant([
    [
        [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]
    ],
    [
        [1.0, 0.0, 0.0, 0.1],
        [0.0, 1.0, 0.0, 0.1],
        [0.0, 0.0, 1.0, 0.1],
        [0.0, 0.0, 0.0, 1.0]
    ]
], dtype=tf.float32)

# image_size: shape [A1, ..., An, 2] representing width and height
# Shape [2, 2] to match batch dimension
image_size = tf.constant([
    [image_width, image_height],
    [image_width, image_height]
], dtype=tf.int32)

# Optional name parameter
name = "test_unproject"

ARGS = {
    "pixel_locations": pixel_locations,
    "clip_to_world_matrix": clip_to_world_matrix,
    "image_size": image_size,
    "name": name
}
Setup block drafted by
a language model, Claude Haiku 4.5, prompt python_fixture_v1, through the batch API
Imports
numpy, tensorflow — 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
uvs_to_pixel_indices Ran
ran on a synthesized input
pmh47/dirt/samples/textured.py
code served (permissive licence) · get_code("5ef71faf0ccb184d")
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 EagerTensor, ndim 3, dtype family float, shape (2, 4, 2) 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 tensorflow as tf
import numpy as np

# Create UV coordinates in the range [0, 1]
# Shape: (2, 4, 2) - batch of 2, 4 UV pairs each
uvs_data = np.array([
    [[0.1, 0.2], [0.5, 0.6], [0.9, 0.3], [0.4, 0.7]],
    [[0.2, 0.4], [0.8, 0.1], [0.3, 0.9], [0.6, 0.5]]
], dtype=np.float32)
uvs = tf.constant(uvs_data)

# Texture shape as a tensor [height, width]
texture_shape = tf.constant([64, 128], dtype=tf.int32)

# Mode for handling UV coordinates
mode = 'repeat'

ARGS = {
    'uvs': uvs,
    'texture_shape': texture_shape,
    'mode': mode
}
Setup block drafted by
a language model, Claude Haiku 4.5, prompt python_fixture_v1, through the batch API
Imports
numpy, tensorflow — import names as written, not package names to install; a name the repository itself provides is not a package
Returned
type EagerTensor, ndim 3, dtype family float, shape (2, 4, 2) — shape (2, 4, 2) 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
rasterise Not yet run
unverified (harvested, no recorded run)
pmh47/dirt/dirt/rasterise_ops.py
code served (permissive licence) · get_code("17c7c71b5bf52d12")
How it ran

no invocation recorded.

rasterise_batch Not yet run
unverified (harvested, no recorded run)
pmh47/dirt/dirt/rasterise_ops.py
code served (permissive licence) · get_code("2a741f1765f6f34f")
How it ran

no invocation recorded.

rasterise_deferred Not yet run
unverified (harvested, no recorded run)
pmh47/dirt/dirt/rasterise_ops.py
code served (permissive licence) · get_code("428d4e92d92066f5")
How it ran

no invocation recorded.

vertex_normals Not yet run
unverified (harvested, no recorded run)
pmh47/dirt/dirt/lighting.py
code served (permissive licence) · get_code("e0f7c9d31116c78b")
How it ran

no invocation recorded.

vertex_normals_pre_split Not yet run
unverified (harvested, no recorded run)
pmh47/dirt/dirt/lighting.py
code served (permissive licence) · get_code("ad162bd8aff4e9cf")
How it ran

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, 0 are also present in other papers whose code we hold (0 with at least one paper dated earlier, 0 the same month, 0 later) and 13 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.

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 a unified framework tackling two problems: class-specific 3D reconstruction from a single image, and generation of new 3D shape samples. These tasks have received considerable attention recently; however, most existing approaches rely on 3D supervision, annotation of 2D images with keypoints or poses, and/or training with multiple views of each object instance. Our framework is very general: it can be trained in similar settings to existing approaches, while also supporting weaker supervision. Importantly, it can be trained purely from 2D images, without pose annotations, and with only a single view per instance. We employ meshes as an output representation, instead of voxels used in most prior work. This allows us to reason over lighting parameters and exploit shading information during training, which previous 2D-supervised methods cannot. Thus, our method can learn to generate and reconstruct concave object classes. We evaluate our approach in various settings, showing that: (i) it learns to disentangle shape from pose and lighting; (ii) using shading in the loss improves performance compared to just silhouettes; (iii) when using a standard single white light, our model outperforms state-of-the-art 2D-supervised methods, both with and without pose supervision, thanks to exploiting shading cues; (iv) performance improves further when using multiple coloured lights, even approaching that of state-of-the-art 3D-supervised methods; (v) shapes produced by our model capture smooth surfaces and fine details better than voxel-based approaches; and (vi) our approach supports concave classes such as bathtubs and sofas, which methods based on silhouettes cannot learn.

For agents

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

get_harvested_code_for_paper("1901.06447")
get_code_for_paper("1901.06447")
have("1901.06447")

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

curl https://syntology.ai/api/ran/1901.06447.json

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

[![Syntology run record](https://syntology.ai/api/ran/1901.06447.svg)](https://syntology.ai/paper/1901.06447)

Connect an agent — have() is free.