SYNTOLOGY HomeExplorerAtlasCodeMethodologyAboutDevelopersFeedPricing
Paper · 1912.03426 · 2019

Self-Supervised 3D Keypoint Learning for Ego-motion Estimation

arXiv · PDF · Open in the Atlas

Code that ran

9 of 9 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 5 have no contract verdict; 3 failures of our instrument; 0 no recorded run), measured 2026-11-21; 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/1912.03426.json · badge: /api/ran/1912.03426.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
TRI-ML/KP3D canonical (classifier label)
classifier basis: census
9 ran · 0 no recorded run
FunctionStatusWhere it lives
construct_K Ran
ran, honoured the contract we drafted
TRI-ML/KP3D/kp3d/geometry/camera.py
code served (permissive licence) · get_code("3a7ae036e8d2457d")
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 (3, 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)

ARGS = {
    "fx": 500.0,
    "fy": 500.0,
    "cx": 320.0,
    "cy": 240.0,
    "dtype": torch.float,
    "device": None,
}
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 (3, 3) — shape (3, 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
image_grid Ran
ran on a synthesized input
TRI-ML/KP3D/kp3d/utils/image.py
code served (permissive licence) · get_code("d462e2f3cd90e4fc")
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 4, dtype family float, shape (2, 3, 4, 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)

ARGS = {
    "B": 2,
    "H": 4,
    "W": 3,
    "dtype": torch.float32,
    "device": "cpu",
    "ones": True,
    "normalized": False,
}
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 4, dtype family float, shape (2, 3, 4, 3) — shape (2, 3, 4, 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
invert_pose Ran
ran on a synthesized input
TRI-ML/KP3D/kp3d/geometry/pose.py
code served (permissive licence) · get_code("a04d16d3da568d50")
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 (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 torch

torch.manual_seed(0)

# Create a batch of 2 homogeneous transformation matrices (4x4)
# Each is a valid SE(3) transformation matrix
batch_size = 2
T01 = torch.eye(4, dtype=torch.float32).repeat(batch_size, 1, 1)

# Modify to create distinct transformations
# Add some rotation and translation
T01[0, :3, :3] = torch.tensor([
    [0.9, -0.436, 0.0],
    [0.436, 0.9, 0.0],
    [0.0, 0.0, 1.0]
], dtype=torch.float32)
T01[0, :3, 3] = torch.tensor([1.0, 2.0, 3.0], dtype=torch.float32)

T01[1, :3, :3] = torch.tensor([
    [0.707, 0.0, 0.707],
    [0.0, 1.0, 0.0],
    [-0.707, 0.0, 0.707]
], dtype=torch.float32)
T01[1, :3, 3] = torch.tensor([0.5, 1.5, 2.5], dtype=torch.float32)

ARGS = {
    "T01": T01
}
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 (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
meshgrid Ran
ran on a synthesized input
TRI-ML/KP3D/kp3d/utils/image.py
code served (permissive licence) · get_code("5b11715c4e4a1c7f")
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 4, 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 = {
    "B": 2,
    "H": 4,
    "W": 3,
    "dtype": torch.float32,
    "device": "cpu",
    "normalized": False,
}
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 4, 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
read_calib_file Ran
ran; our contract draft was wrong, not the code
TRI-ML/KP3D/kp3d/datasets/kitti_odometry_dataset.py
code served (permissive licence) · get_code("3c91deaa8c673fd6")
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 tempfile
import os

# Create a temporary calibration file with valid content
temp_dir = tempfile.mkdtemp()
filepath = os.path.join(temp_dir, "calibration.txt")

with open(filepath, 'w') as f:
    f.write("P0: 721.5377 0.0 609.5593 44.85728 0.0 721.5377 172.854 0.2163 0.0 0.0 1.0 0.002677\n")
    f.write("P1: 721.5377 0.0 609.5593 -40.37714 0.0 721.5377 172.854 0.2163 0.0 0.0 1.0 0.002677\n")
    f.write("R0_rect: 0.9999239 0.0105192 -0.0059216 -0.0104963 0.9999277 0.0063487 0.0059393 -0.0063261 0.9999622\n")
    f.write("Tr_velo_to_cam: 0.07967 -0.99977 -0.00701 -0.07601 0.01710 -0.00665 0.99982 -0.11002 -0.99972 -0.07973 0.01625 -0.05369\n")

ARGS = {
    "filepath": filepath
}
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 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
scale_intrinsics Ran
ran; the contract check could not run (RAISES, DEP_MISSING or TIMEOUT)
TRI-ML/KP3D/kp3d/geometry/camera.py
code served (permissive licence) · get_code("b9ff60bfcb5831da")
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 (2, 3, 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
import numpy as np

torch.manual_seed(0)

# Test case 1: torch tensor with shape (3, 3) - single intrinsic matrix
K_single = torch.tensor([
    [500.0, 0.0, 320.0],
    [0.0, 500.0, 240.0],
    [0.0, 0.0, 1.0]
])

# Test case 2: torch tensor with shape (2, 3, 3) - batch of intrinsic matrices
K_batch = torch.tensor([
    [[500.0, 0.0, 320.0],
     [0.0, 500.0, 240.0],
     [0.0, 0.0, 1.0]],
    [[600.0, 0.0, 640.0],
     [0.0, 600.0, 480.0],
     [0.0, 0.0, 1.0]]
])

x_scale = 0.5
y_scale = 0.75

# Use batch version for testing
ARGS = {
    "K": K_batch.clone(),
    "x_scale": x_scale,
    "y_scale": y_scale
}
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 Tensor, ndim 3, dtype family float, shape (2, 3, 3) — shape (2, 3, 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
select_k_best Ran
ran on a synthesized input
TRI-ML/KP3D/kp3d/evaluation/descriptor_evaluation.py
code served (permissive licence) · get_code("5487d34d4f82c3f1")
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

# Construct valid arguments for select_k_best
num_points = 5
k_value = 3

# Create points array with shape (num_points, 3)
# Last column is probability
points = np.array([
    [10.0, 20.0, 0.5],
    [30.0, 40.0, 0.9],
    [50.0, 60.0, 0.3],
    [70.0, 80.0, 0.8],
    [90.0, 100.0, 0.6]
], dtype=np.float32)

# Create descriptors array with shape (num_points, 256)
descriptors = np.random.randn(num_points, 256).astype(np.float32)

ARGS = {
    "points": points,
    "descriptors": descriptors,
    "k": k_value
}
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
to_gray_normalized Ran
ran on a synthesized input
TRI-ML/KP3D/kp3d/utils/image.py
code served (permissive licence) · get_code("1359c08a986ae150")
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 4, dtype family float, shape (2, 1, 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 batch of images
# Shape: (batch_size=2, channels=3, height=4, width=4)
images = torch.randn(2, 3, 4, 4)

ARGS = {
    "images": images
}
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 4, dtype family float, shape (2, 1, 4, 4) — shape (2, 1, 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
upsample Ran
ran; the contract check could not run (RAISES, DEP_MISSING or TIMEOUT)
TRI-ML/KP3D/kp3d/networks/keypoint_resnet.py
pointer only (licence: MIT); the stored copy get_code serves, from leofansq/CBR, is pointer only (licence: NONE) · get_code("a739072b350093e7")
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 leofansq/CBR, 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.

Also present in other papers' code

Of 9 samples lifted from this paper's linked repositories, 9 carry a function identity and 0 do not. Of the 9 distinct functions with an identity, 4 are also present in other papers whose code we hold (0 with at least one paper dated earlier, 0 the same month, 4 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
upsample 8 other papers
8 later
2002.08394 (2020-02)
MonoLayout: Amodal scene layout from a single image
2103.14201 (2021-03)
Image2Reverb: Cross-Modal Reverb Impulse Response Synthesis
2108.08829 (2021-08)
Fine-grained Semantics-aware Representation Enhancement for Self-supervised Monocular Depth Estimation
and 5 more not named here
construct_K 2 other papers
2 later
2002.12319 (2020-02)
Semantically-Guided Representation Learning for Self-Supervised Monocular Depth
2103.16694 (2021-03)
Geometric Unsupervised Domain Adaptation for Semantic Segmentation
read_calib_file 2 other papers
2 later
2403.11270 (2024-03)
Bilateral Propagation Network for Depth Completion
2601.21291 (2026-01)
Gaussian Belief Propagation Network for Depth Completion
scale_intrinsics 2 other papers
2 later
2002.12319 (2020-02)
Semantically-Guided Representation Learning for Self-Supervised Monocular Depth
2103.16694 (2021-03)
Geometric Unsupervised Domain Adaptation for Semantic 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 2026-11-21.

Repositories linked to this paper

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

Abstract

Detecting and matching robust viewpoint-invariant keypoints is critical for visual SLAM and Structure-from-Motion. State-of-the-art learning-based methods generate training samples via homography adaptation to create 2D synthetic views with known keypoint matches from a single image. This approach, however, does not generalize to non-planar 3D scenes with illumination variations commonly seen in real-world videos. In this work, we propose self-supervised learning of depth-aware keypoints directly from unlabeled videos. We jointly learn keypoint and depth estimation networks by combining appearance and geometric matching via a differentiable structure-from-motion module based on Procrustean residual pose correction. We describe how our self-supervised keypoints can be integrated into state-of-the-art visual odometry frameworks for robust and accurate ego-motion estimation of autonomous vehicles in real-world conditions.

For agents

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

get_harvested_code_for_paper("1912.03426")
get_code_for_paper("1912.03426")
have("1912.03426")

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

curl https://syntology.ai/api/ran/1912.03426.json

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

[![Syntology run record](https://syntology.ai/api/ran/1912.03426.svg)](https://syntology.ai/paper/1912.03426)

Connect an agent — have() is free.