SYNTOLOGY HomeExplorerAtlasCodeMethodologyAboutDevelopersFeedPricing
Paper · 2203.14517 · CVPR · 2022

REGTR: End-to-end Point Cloud Correspondences with Transformers

Hee Lee, Zi Jian, Yew Gim

arXiv · PDF · Open in the Atlas

Code that ran

10 of 16 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 10 have no contract verdict; 0 failures of our instrument; 6 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/2203.14517.json · badge: /api/ran/2203.14517.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
yewzijian/RegTR canonical (classifier label)
classifier basis not recorded on the link
10 ran · 6 no recorded run
FunctionStatusWhere it lives
all_isfinite Ran
ran on a synthesized input
yewzijian/RegTR/src/cvhelpers/torch_helpers.py
code served (permissive licence) · get_code("b16bf332b47d6d18")
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 bool, 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

torch.manual_seed(0)

# Test case 1: A simple finite tensor
finite_tensor = torch.randn(2, 4)

# Test case 2: A list of finite tensors
list_of_tensors = [torch.randn(2, 4), torch.randn(3, 5)]

# Test case 3: A nested list of tensors
nested_list = [torch.randn(2, 4), [torch.randn(3, 5), torch.randn(2, 3)]]

# Test case 4: A tensor with inf values
tensor_with_inf = torch.tensor([[1.0, float('inf')], [2.0, 3.0]])

# Test case 5: A tensor with nan values
tensor_with_nan = torch.tensor([[1.0, float('nan')], [2.0, 3.0]])

# Test case 6: A list containing a tensor with inf
list_with_inf = [torch.randn(2, 4), tensor_with_inf]

# Use the finite tensor for the main test case
ARGS = {
    "x": finite_tensor
}
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 bool, 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
collate_pair Ran
ran on a synthesized input
yewzijian/RegTR/src/data_loaders/collate_functions.py
code served (permissive licence) · get_code("6c01815b97cb00fc")
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 torch

torch.manual_seed(0)

# Create sample data items for the batch
# Each item is a dict with keys that collate_pair expects
sample_item_1 = {
    'src_xyz': torch.randn(5, 3),  # 5 points, 3D coordinates
    'tgt_xyz': torch.randn(7, 3),  # 7 points, 3D coordinates (different size)
    'tgt_raw': torch.randn(7, 4),  # raw target data
    'src_overlap': torch.randn(5),
    'tgt_overlap': torch.randn(7),
    'correspondences': torch.tensor([[0, 1], [1, 2]]),
    'src_path': 'src_path_1.txt',
    'tgt_path': 'tgt_path_1.txt',
    'idx': 0,
    'pose': torch.eye(4),  # 4x4 pose matrix
    'overlap_p': 0.5,
}

sample_item_2 = {
    'src_xyz': torch.randn(6, 3),  # different number of points
    'tgt_xyz': torch.randn(8, 3),
    'tgt_raw': torch.randn(8, 4),
    'src_overlap': torch.randn(6),
    'tgt_overlap': torch.randn(8),
    'correspondences': torch.tensor([[0, 2], [2, 3], [3, 4]]),
    'src_path': 'src_path_2.txt',
    'tgt_path': 'tgt_path_2.txt',
    'idx': 1,
    'pose': torch.eye(4) * 2,
    'overlap_p': 0.7,
}

list_data = [sample_item_1, sample_item_2]

ARGS = {
    'list_data': list_data,
}
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 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
compute_rte Ran
ran on a synthesized input
yewzijian/RegTR/src/benchmark/benchmark_3dmatch.py
code served (permissive licence) · get_code("e5d44638e7b011b6")
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 float64, ndim 0, 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

# The function compute_rte takes two parameters: t and t_est
# Both should be numpy arrays representing translation vectors
# The function computes the Euclidean norm of their difference

# Create small translation vectors (3D translations are common in computer vision)
t = np.array([1.0, 2.0, 3.0])
t_est = np.array([1.1, 2.1, 2.9])

ARGS = {
    "t": t,
    "t_est": t_est,
}
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 float64, ndim 0, dtype family float — no output shape is stated: observe_signature.py recorded float ndim 0, which disagrees with the fixture run
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
dcm2euler Ran
ran on a synthesized input
yewzijian/RegTR/src/benchmark/benchmark_modelnet.py
code served (permissive licence) · get_code("5b694f7ba87b048f")
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 (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 numpy as np
from scipy.spatial.transform import Rotation

# Create valid rotation matrices for batch size 2
# A rotation matrix must be orthogonal (R^T * R = I) and have det(R) = 1
# We'll use scipy.spatial.transform.Rotation to generate valid ones

batch_size = 2

# Generate random valid rotation matrices using Rotation
rotations = Rotation.random(num=batch_size, random_state=42)
mats = rotations.as_matrix()  # Shape: (2, 3, 3)

# Verify the matrices are valid rotation matrices
assert mats.shape == (batch_size, 3, 3), f"Expected shape (2, 3, 3), got {mats.shape}"
for i in range(batch_size):
    # Check orthogonality: R^T * R should be identity
    check = np.allclose(mats[i] @ mats[i].T, np.eye(3), atol=1e-6)
    assert check, f"Matrix {i} is not orthogonal"
    # Check determinant is 1
    det = np.linalg.det(mats[i])
    assert np.isclose(det, 1.0, atol=1e-6), f"Matrix {i} has det={det}"

ARGS = {
    'mats': mats,
    'seq': 'zyx',
    'degrees': True
}
Setup block drafted by
a language model, Claude Haiku 4.5, prompt python_fixture_v1, through the batch API
Imports
numpy, scipy — 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 (2, 3) — shape (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
pretty_time_delta Ran
ran on a synthesized input
yewzijian/RegTR/src/cvhelpers/misc.py
code served (permissive licence) · get_code("519457a60a677d35")
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 str. 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 numpy as np

# The function pretty_time_delta takes a single parameter: seconds (a numeric value)
# It converts this to days, hours, minutes, and seconds for pretty printing
# No complex objects needed - just numeric values

# Test with various time deltas
ARGS = {
    "seconds": 93784  # 1 day, 2 hours, 3 minutes, 4 seconds
}
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 str — a str output has no array shape of its own; none is stated — the instrument coerces any output to an array; for a str 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
rotation_error Ran
ran on a synthesized input
yewzijian/RegTR/src/benchmark/benchmark_predator.py
code served (permissive licence) · get_code("3344db52a0beedde")
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, 1) 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 math

torch.manual_seed(0)

# Create two batches of 3x3 rotation matrices
batch_size = 2

# R1: estimated rotation matrices [2, 3, 3]
R1 = torch.randn(batch_size, 3, 3)
# Orthogonalize to make them valid rotation matrices
U, S, V = torch.svd(R1)
R1 = torch.matmul(U, V.transpose(1, 2))

# R2: ground truth rotation matrices [2, 3, 3]
R2 = torch.randn(batch_size, 3, 3)
# Orthogonalize to make them valid rotation matrices
U, S, V = torch.svd(R2)
R2 = torch.matmul(U, V.transpose(1, 2))

ARGS = {
    "R1": R1,
    "R2": R2,
}
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, 1) — shape (2, 1) 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
summarize_metrics Ran
ran on a synthesized input
yewzijian/RegTR/src/benchmark/benchmark_modelnet.py
code served (permissive licence) · get_code("14b119a1f275889c")
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

ARGS = {
    'metrics': {
        'loss_mse': np.array([0.5, 0.3, 0.8, 0.2]),
        'err_count': np.array([1, 2, 3, 4]),
        'accuracy': np.array([0.95, 0.92, 0.98, 0.91]),
    }
}
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
to_numpy Ran
ran on a synthesized input
yewzijian/RegTR/src/cvhelpers/torch_helpers.py
code served (permissive licence) · get_code("2fa7b4714e1f9d24")
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 (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
import numpy as np

torch.manual_seed(0)

# Test case 1: torch.Tensor
tensor_arg = torch.randn(2, 4)

# Test case 2: np.ndarray
numpy_arg = np.random.randn(3, 5)

# Test case 3: List containing mixed types
list_arg = [
    torch.randn(2, 4),
    np.array([1, 2, 3]),
    "string_value",
    None,
    [torch.randn(1, 2), np.array([4, 5])]
]

# Define ARGS - testing the function with a torch.Tensor as the primary case
ARGS = {
    "tensor": tensor_arg
}
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 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
translation_error Ran
ran on a synthesized input
yewzijian/RegTR/src/benchmark/benchmark_predator.py
code served (permissive licence) · get_code("81b265a618542b54")
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 1, dtype family float, shape (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 torch

torch.manual_seed(0)

# The function expects:
# t1: estimated translation vectors [b,3,1]
# t2: ground truth translation vectors [b,3,1]
# Returns: translation error in meters [b,1]

# Create small batch of translation vectors
batch_size = 2
t1 = torch.randn(batch_size, 3, 1)
t2 = torch.randn(batch_size, 3, 1)

ARGS = {
    "t1": t1,
    "t2": t2,
}
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 1, dtype family float, shape (2,) — shape (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
uniform_2_sphere Ran
ran on a synthesized input
yewzijian/RegTR/src/data_loaders/modelnet_transforms.py
code served (permissive licence) · get_code("a020832e85dfd32f")
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 (5, 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

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

# Define ARGS dict with all parameters
ARGS = {
    "num": 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 (5, 3) — shape (5, 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
all_to_device Not yet run
unverified (harvested, no recorded run)
yewzijian/RegTR/src/cvhelpers/torch_helpers.py
code served (permissive licence) · get_code("e7aaf57404c17d96")
How it ran

no invocation recorded.

get_test_datasets Not yet run
unverified (harvested, no recorded run)
yewzijian/RegTR/src/data_loaders/modelnet.py
code served (permissive licence) · get_code("43be91adb1cc94fa")
How it ran

no invocation recorded.

get_train_datasets Not yet run
unverified (harvested, no recorded run)
yewzijian/RegTR/src/data_loaders/modelnet.py
code served (permissive licence) · get_code("a149b629c86cfa52")
How it ran

no invocation recorded.

get_transforms Not yet run
unverified (harvested, no recorded run)
yewzijian/RegTR/src/data_loaders/modelnet.py
code served (permissive licence) · get_code("e659b44bcc28934a")
How it ran

no invocation recorded.

read_trajectory Not yet run
unverified (harvested, no recorded run)
yewzijian/RegTR/src/benchmark/benchmark_3dmatch.py
code served (permissive licence) · get_code("48f96ab357b5c0e0")
How it ran

no invocation recorded.

read_trajectory_info Not yet run
unverified (harvested, no recorded run)
yewzijian/RegTR/src/benchmark/benchmark_3dmatch.py
code served (permissive licence) · get_code("59dff5c1fef3d449")
How it ran

no invocation recorded.

Also present in other papers' code

Of 16 samples lifted from this paper's linked repositories, 16 carry a function identity and 0 do not. Of the 16 distinct functions with an identity, 1 are also present in other papers whose code we hold (1 with at least one paper dated earlier, 0 the same month, 0 later) and 15 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
uniform_2_sphere 2 other papers
1 dated earlier, 1 same year, order not determinable
2003.13479 (2020-03)
RPM-Net: Robust Point Matching using Learned Features
aaai_20189 (2022 (year only))
FINet: Dual Branches Feature Interaction for Partial-to-Partial Point Cloud Registration

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

Despite recent success in incorporating learning into point cloud registration, many works focus on learning feature descriptors and continue to rely on nearest-neighbor feature matching and outlier filtering through RANSAC to obtain the final set of correspondences for pose estimation. In this work, we conjecture that attention mechanisms can replace the role of explicit feature matching and RANSAC, and thus propose an end-to-end framework to directly predict the final set of correspondences. We use a network architecture consisting primarily of transformer layers containing self and cross attentions, and train it to predict the probability each point lies in the overlapping region and its corresponding position in the other point cloud. The required rigid transformation can then be estimated directly from the predicted correspondences without further post-processing. Despite its simplicity, our approach achieves state-of-the-art performance on 3DMatch and ModelNet benchmarks. Our source code can be found at https://github.com/yewzijian/RegTR.

For agents

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

get_harvested_code_for_paper("2203.14517")
get_code_for_paper("2203.14517")
have("2203.14517")

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

curl https://syntology.ai/api/ran/2203.14517.json

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

[![Syntology run record](https://syntology.ai/api/ran/2203.14517.svg)](https://syntology.ai/paper/2203.14517)

Connect an agent — have() is free.