SYNTOLOGY HomeExplorerAtlasCodeMethodologyAboutDevelopersFeedPricing
Paper · 2003.13479 · 2020

RPM-Net: Robust Point Matching using Learned Features

arXiv · PDF · Open in the Atlas

Code that ran

26 of 31 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 19 have no contract verdict; 6 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/2003.13479.json · badge: /api/ran/2003.13479.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/RPMNet canonical (classifier label)
classifier basis: census
15 ran · 2 no recorded run
vinits5/masknet community (classifier label)
classifier basis not recorded on the link
11 ran · 3 no recorded run
FunctionStatusWhere it lives
add_outliers Ran
ran on a synthesized input
vinits5/masknet/evaluation/create_statistical_data.py
pointer only (licence: MIT) · get_code("5706ca4fef0f2696")
How it ran

The body of this sample is withheld: MIT, as GitHub detects it on this paper's repository, vinits5/masknet, is in the permissive set, but a licence decision keeps the body a pointer: the reason for that decision is not recorded with it. 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.

angle_difference Ran
ran; the contract check could not run (RAISES, DEP_MISSING or TIMEOUT)
yewzijian/RPMNet/src/models/pointnet_util.py
code served (permissive licence) · get_code("e6531f546e55f6af")
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, 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)

# src: [B, N, C] - batch of N source points with C channels
# dst: [B, M, C] - batch of M target points with C channels
# The function assumes l2-normalized unit vectors and computes angles between them

B = 2  # batch size
N = 3  # number of source points
M = 4  # number of target points
C = 5  # channel/feature dimension

# Create l2-normalized unit vectors for src and dst
src = torch.randn(B, N, C)
src = src / torch.norm(src, dim=2, keepdim=True)  # normalize to unit length

dst = torch.randn(B, M, C)
dst = dst / torch.norm(dst, dim=2, keepdim=True)  # normalize to unit length

ARGS = {
    "src": src,
    "dst": dst,
}
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, 3, 4) — shape (2, 3, 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
classification_loss Ran
ran on a synthesized input
vinits5/masknet/learning3d/losses/classification.py
pointer only (licence: MIT) · get_code("78cd571bf9c38cc0")
How it ran

The body of this sample is withheld: MIT, as GitHub detects it on this paper's repository, vinits5/masknet, is in the permissive set, but a licence decision keeps the body a pointer: the reason for that decision is not recorded with it. 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.

concatenate Ran
ran on a synthesized input
yewzijian/RPMNet/src/common/math/se3.py
code served (permissive licence) · get_code("8a632dcefddc5fc1")
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 (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 numpy as np

# Create two SE3 transform matrices
# SE3 transforms are 4x4 matrices (or 3x4 for homogeneous representation)
# Format: [R | t] where R is 3x3 rotation, t is 3x1 translation
#         [0 | 1] (optional last row)

# First transform: 4x4 matrix
a = np.array([
    [1.0, 0.0, 0.0, 1.0],
    [0.0, 1.0, 0.0, 2.0],
    [0.0, 0.0, 1.0, 3.0],
    [0.0, 0.0, 0.0, 1.0]
])

# Second transform: 4x4 matrix
b = np.array([
    [1.0, 0.0, 0.0, 0.5],
    [0.0, 1.0, 0.0, 0.5],
    [0.0, 0.0, 1.0, 0.5],
    [0.0, 0.0, 0.0, 1.0]
])

ARGS = {
    "a": a,
    "b": b
}
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 (4, 4) — shape (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
concatenate Ran
ran on a synthesized input
yewzijian/RPMNet/src/common/math_torch/se3.py
code served (permissive licence) · get_code("5eef8614474cef47")
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, 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)

# SE3 transforms are represented as (B, 3, 4) matrices
# where the first 3x3 block is rotation and the last column is translation
# We need two such transforms to concatenate

batch_size = 2

# Create a: (B, 3, 4) SE3 transform
# 3x3 rotation block + 3x1 translation vector
a = torch.randn(batch_size, 3, 4)

# Create b: (B, 3, 4) SE3 transform  
# 3x3 rotation block + 3x1 translation vector
b = torch.randn(batch_size, 3, 4)

ARGS = {
    "a": a,
    "b": b
}
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, 3, 4) — shape (2, 3, 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
farthest_subsample_points Ran
ran on a synthesized input
vinits5/masknet/evaluation/create_statistical_data.py
pointer only (licence: MIT) · get_code("1f8e2a3d2a65c39a")
How it ran

The body of this sample is withheld: MIT, as GitHub detects it on this paper's repository, vinits5/masknet, is in the permissive set, but a licence decision keeps the body a pointer: the reason for that decision is not recorded with it. 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.

find_pretrained_path Ran
ran on a synthesized input
vinits5/masknet/registration.py
pointer only (licence: MIT) · get_code("b0f6aff69d22bcdd")
How it ran

The body of this sample is withheld: MIT, as GitHub detects it on this paper's repository, vinits5/masknet, is in the permissive set, but a licence decision keeps the body a pointer: the reason for that decision is not recorded with it. 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.

flip_geometries Ran
ran on a synthesized input
vinits5/masknet/3dmatch/plot_figures.py
pointer only (licence: MIT) · get_code("f41a4accdacef815")
How it ran

The body of this sample is withheld: MIT, as GitHub detects it on this paper's repository, vinits5/masknet, is in the permissive set, but a licence decision keeps the body a pointer: the reason for that decision is not recorded with it. 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.

get_angle_data Ran
ran on a synthesized input
vinits5/masknet/evaluation/plot_results.py
pointer only (licence: MIT) · get_code("39959096693129c5")
How it ran

The body of this sample is withheld: MIT, as GitHub detects it on this paper's repository, vinits5/masknet, is in the permissive set, but a licence decision keeps the body a pointer: the reason for that decision is not recorded with it. 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.

get_files Ran
ran on a synthesized input
vinits5/masknet/3dmatch/make_video.py
pointer only (licence: MIT) · get_code("a45813b942a85200")
How it ran

The body of this sample is withheld: MIT, as GitHub detects it on this paper's repository, vinits5/masknet, is in the permissive set, but a licence decision keeps the body a pointer: the reason for that decision is not recorded with it. 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.

get_noise_data Ran
ran on a synthesized input
vinits5/masknet/evaluation/plot_results.py
pointer only (licence: MIT) · get_code("fbc79bc476871f2f")
How it ran

The body of this sample is withheld: MIT, as GitHub detects it on this paper's repository, vinits5/masknet, is in the permissive set, but a licence decision keeps the body a pointer: the reason for that decision is not recorded with it. 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.

get_postpool Ran
ran; our contract draft was wrong, not the code
yewzijian/RPMNet/src/models/feature_nets.py
code served (permissive licence) · get_code("552ad03917659f63")
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 Sequential. 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_dim": 64,
    "out_dim": 32,
}
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 Sequential — a Sequential output has no array shape of its own; none is stated — the instrument coerces any output to an array; for a Sequential output the recorded ndim 1 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
get_prepool Ran
ran; our contract draft was wrong, not the code
yewzijian/RPMNet/src/models/feature_nets.py
code served (permissive licence) · get_code("8778312931854a3b")
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 Sequential. 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_dim": 3,
    "out_dim": 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 Sequential — a Sequential output has no array shape of its own; none is stated — the instrument coerces any output to an array; for a Sequential output the recorded ndim 1 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
identity Ran
ran on a synthesized input
yewzijian/RPMNet/src/common/math_torch/se3.py
code served (permissive licence) · get_code("89304006422410bd")
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, 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)

ARGS = {
    "batch_size": 2
}
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, 3, 4) — shape (2, 3, 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
index_points Ran
ran; the contract check could not run (RAISES, DEP_MISSING or TIMEOUT)
yewzijian/RPMNet/src/models/pointnet_util.py
code served (permissive licence) · get_code("0188255c8db5f0c5")
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, 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)

# points: [B, N, C] - batch size 2, N points 5, C channels 3
points = torch.randn(2, 5, 3)

# idx: [B, S] - batch size 2, S samples 4
# indices should be valid (0 to N-1, so 0 to 4)
idx = torch.tensor([[0, 2, 1, 3], [4, 1, 2, 0]], dtype=torch.long)

ARGS = {
    "points": points,
    "idx": idx
}
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, 3) — shape (2, 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
inverse Ran
ran on a synthesized input
yewzijian/RPMNet/src/common/math/se3.py
code served (permissive licence) · get_code("caea4df4266011b9")
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 (3, 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 numpy as np

# Create a valid SE3 transform (3x4 form without homogeneous row)
# A valid SE3 is [R | t] where R is a 3x3 rotation matrix and t is 3x1 translation
rotation_matrix = np.array([
    [1.0, 0.0, 0.0],
    [0.0, 1.0, 0.0],
    [0.0, 0.0, 1.0]
], dtype=np.float64)

translation_vector = np.array([[1.0], [2.0], [3.0]], dtype=np.float64)

g_3x4 = np.concatenate([rotation_matrix, translation_vector], axis=1)

ARGS = {
    "g": g_3x4
}
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 (3, 4) — shape (3, 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
inverse Ran
ran on a synthesized input
yewzijian/RPMNet/src/common/math_torch/se3.py
code served (permissive licence) · get_code("c3f07450b8d24c41")
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, 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 SE3 transforms
# Shape: (B, 3/4, 4) where B=2 (batch size)
# We'll use (B, 4, 4) format which is standard for SE3 matrices
# The function extracts g[..., 0:3, 0:3] for rotation and g[..., 0:3, 3] for translation

B = 2
g = torch.zeros(B, 4, 4)

# Create valid SE3 transforms (identity + small perturbation)
for i in range(B):
    # Rotation part (top-left 3x3) - use orthogonal matrices
    g[i, 0:3, 0:3] = torch.eye(3) + 0.01 * torch.randn(3, 3)
    # Make approximately orthogonal via QR decomposition
    Q, R = torch.linalg.qr(g[i, 0:3, 0:3])
    g[i, 0:3, 0:3] = Q
    
    # Translation part (top-right 3x1)
    g[i, 0:3, 3] = torch.randn(3)
    
    # Bottom row for homogeneous coordinates
    g[i, 3, :] = torch.tensor([0.0, 0.0, 0.0, 1.0])

ARGS = {
    "g": g
}
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, 3, 4) — shape (2, 3, 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
pc2points Ran
ran on a synthesized input
vinits5/masknet/evaluation/evaluate_stats.py
pointer only (licence: MIT) · get_code("9e2e626d60f3368d")
How it ran

The body of this sample is withheld: MIT, as GitHub detects it on this paper's repository, vinits5/masknet, is in the permissive set, but a licence decision keeps the body a pointer: the reason for that decision is not recorded with it. 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.

read_all_files Ran
ran on a synthesized input
vinits5/masknet/evaluation/plot_results.py
pointer only (licence: MIT) · get_code("35d8b1f2dc9d527e")
How it ran

The body of this sample is withheld: MIT, as GitHub detects it on this paper's repository, vinits5/masknet, is in the permissive set, but a licence decision keeps the body a pointer: the reason for that decision is not recorded with it. 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.

read_data Ran
ran on a synthesized input
vinits5/masknet/3dmatch/plot_figures.py
pointer only (licence: MIT) · get_code("a326f2a06da4eba0")
How it ran

The body of this sample is withheld: MIT, as GitHub detects it on this paper's repository, vinits5/masknet, is in the permissive set, but a licence decision keeps the body a pointer: the reason for that decision is not recorded with it. 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.

sinkhorn Ran
ran; the contract check could not run (RAISES, DEP_MISSING or TIMEOUT)
yewzijian/RPMNet/src/models/rpmnet.py
code served (permissive licence) · get_code("d9502fe2fe76bf06")
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, 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 torch.nn as nn

torch.manual_seed(0)

# Construct log_alpha: (B, J, K) where B=batch, J=2, K=3
# Using small distinct dimensions for clarity
log_alpha = torch.randn(2, 4, 3)

# n_iters: number of normalization iterations
n_iters = 5

# slack: whether to include slack row and column
slack = True

# eps: early termination threshold (negative to disable)
eps = -1

ARGS = {
    "log_alpha": log_alpha,
    "n_iters": n_iters,
    "slack": slack,
    "eps": eps,
}
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, 3) — shape (2, 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
square_distance Ran
ran; the contract check could not run (RAISES, DEP_MISSING or TIMEOUT)
yewzijian/RPMNet/src/models/pointnet_util.py
pointer only (licence: MIT); the stored copy get_code serves, from fukexue/RGM, is pointer only (licence: NONE) · get_code("4db73dd57b5d6162")
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 fukexue/RGM, 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.

to_numpy Ran
ran, honoured the contract we drafted
yewzijian/RPMNet/src/common/torch.py
pointer only (licence: MIT); the stored copy get_code serves, from YasiiKB/R3PM-Net, is pointer only (licence: NONE) · get_code("a2a91d3eb9618070")
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 YasiiKB/R3PM-Net, 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.

transform Ran
ran on a synthesized input
yewzijian/RPMNet/src/common/math/se3.py
code served (permissive licence) · get_code("9658f194c41b4f13")
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 (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 numpy as np

# Construct SE3 transformation matrix: shape (3, 4) or (B, 3, 4)
# SE3 format: [R | t] where R is 3x3 rotation, t is 3x1 translation
g_single = np.eye(4)[:3, :]  # shape (3, 4)
g_batch = np.array([np.eye(4)[:3, :], np.eye(4)[:3, :]])  # shape (2, 3, 4)

# Construct points to transform: shape (N, 3) or (B, N, 3)
pts_single = np.array([
    [1.0, 2.0, 3.0],
    [4.0, 5.0, 6.0],
    [7.0, 8.0, 9.0],
    [10.0, 11.0, 12.0]
])  # shape (4, 3)

pts_batch = np.array([
    [[1.0, 2.0, 3.0], [4.0, 5.0, 6.0]],
    [[7.0, 8.0, 9.0], [10.0, 11.0, 12.0]]
])  # shape (2, 2, 3)

# Use single (non-batched) case
ARGS = {
    "g": g_single,
    "pts": pts_single
}
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 (4, 3) — shape (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
transform Ran
ran on a synthesized input
yewzijian/RPMNet/src/common/math/so3.py
code served (permissive licence) · get_code("01535a12b106422c")
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

# Create a valid SO3 rotation matrix (3x3)
# Using a simple rotation matrix around the z-axis
angle = np.pi / 4
g = np.array([
    [np.cos(angle), -np.sin(angle), 0],
    [np.sin(angle), np.cos(angle), 0],
    [0, 0, 1]
])

# Create points to transform (N=2 points, 3D coordinates)
pts = np.array([
    [1.0, 0.0, 0.0],
    [0.0, 1.0, 0.0]
])

ARGS = {
    "g": g,
    "pts": pts
}
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 (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
uniform_2_sphere Ran
ran on a synthesized input
yewzijian/RPMNet/src/common/math/random.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
chamfer_distance Not yet run
unverified (harvested, no recorded run)
vinits5/masknet/learning3d/losses/chamfer_distance.py
pointer only (licence: MIT) · get_code("1adcfbd14aa61be6")
How it ran

The body of this sample is withheld: MIT, as GitHub detects it on this paper's repository, vinits5/masknet, is in the permissive set, but a licence decision keeps the body a pointer: the reason for that decision is not recorded with it. 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. No invocation recorded.

compute_rigid_transform Not yet run
unverified (harvested, no recorded run)
yewzijian/RPMNet/src/models/rpmnet.py
code served (permissive licence) · get_code("d79505bf2d1d5cb9")
How it ran

no invocation recorded.

dcm2euler Not yet run
unverified (harvested, no recorded run)
yewzijian/RPMNet/src/common/math/so3.py
code served (permissive licence) · get_code("7944c9b1c74881cb")
How it ran

no invocation recorded.

emd Not yet run
unverified (harvested, no recorded run)
vinits5/masknet/learning3d/losses/emd.py
pointer only (licence: MIT) · get_code("b83c09e9a41febed")
How it ran

The body of this sample is withheld: MIT, as GitHub detects it on this paper's repository, vinits5/masknet, is in the permissive set, but a licence decision keeps the body a pointer: the reason for that decision is not recorded with it. 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. No invocation recorded.

selectROI Not yet run
unverified (harvested, no recorded run)
vinits5/masknet/3dmatch/make_video.py
pointer only (licence: MIT) · get_code("5ae801ea174c3f2d")
How it ran

The body of this sample is withheld: MIT, as GitHub detects it on this paper's repository, vinits5/masknet, is in the permissive set, but a licence decision keeps the body a pointer: the reason for that decision is not recorded with it. 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. No invocation recorded.

Also present in other papers' code

Of 31 samples lifted from this paper's linked repositories, 31 carry a function identity and 0 do not. Of the 31 distinct functions with an identity, 9 are also present in other papers whose code we hold (1 with at least one paper dated earlier, 0 the same month, 9 later) and 22 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
index_points 42 other papers
2 dated earlier, 40 later · 6 only as an assembly brick
1706.02413 (2017-06)
PointNet++: Deep Hierarchical Feature Learning on Point Sets in a Metric Space
1811.07246 (2018-11)
PointConv: Deep Convolutional Networks on 3D Point Clouds
2004.10904 (2020-04)
Through the Looking Glass: Neural 3D Reconstruction of Transparent Shapes
and 39 more not named here
square_distance 3 other papers
3 later
2103.04256 (2021-03)
Robust Point Cloud Registration Framework Based on Deep Graph Matching
2310.17359 (2023-10)
SE(3) Diffusion Model-based Point Cloud Registration for Robust 6D Object Pose Estimation
2604.05060 (2026-04)
R3PM-Net: Real-time, Robust, Real-world Point Matching Network
angle_difference 2 other papers
2 later
2310.17359 (2023-10)
SE(3) Diffusion Model-based Point Cloud Registration for Robust 6D Object Pose Estimation
2604.05060 (2026-04)
R3PM-Net: Real-time, Robust, Real-world Point Matching Network
uniform_2_sphere 2 other papers
2 later
aaai_20189 (2022 (year only))
FINet: Dual Branches Feature Interaction for Partial-to-Partial Point Cloud Registration
2203.14517 (2022-03)
REGTR: End-to-end Point Cloud Correspondences with Transformers
compute_rigid_transform 1 other paper
1 later
2604.05060 (2026-04)
R3PM-Net: Real-time, Robust, Real-world Point Matching Network
get_postpool 1 other paper
1 later
2310.17359 (2023-10)
SE(3) Diffusion Model-based Point Cloud Registration for Robust 6D Object Pose Estimation
get_prepool 1 other paper
1 later
2310.17359 (2023-10)
SE(3) Diffusion Model-based Point Cloud Registration for Robust 6D Object Pose Estimation
sinkhorn 1 other paper
1 later
2604.05060 (2026-04)
R3PM-Net: Real-time, Robust, Real-world Point Matching Network
to_numpy 1 other paper
1 later
2604.05060 (2026-04)
R3PM-Net: Real-time, Robust, Real-world Point Matching Network

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

Iterative Closest Point (ICP) solves the rigid point cloud registration problem iteratively in two steps: (1) make hard assignments of spatially closest point correspondences, and then (2) find the least-squares rigid transformation. The hard assignments of closest point correspondences based on spatial distances are sensitive to the initial rigid transformation and noisy/outlier points, which often cause ICP to converge to wrong local minima. In this paper, we propose the RPM-Net -- a less sensitive to initialization and more robust deep learning-based approach for rigid point cloud registration. To this end, our network uses the differentiable Sinkhorn layer and annealing to get soft assignments of point correspondences from hybrid features learned from both spatial coordinates and local geometry. To further improve registration performance, we introduce a secondary network to predict optimal annealing parameters. Unlike some existing methods, our RPM-Net handles missing correspondences and point clouds with partial visibility. Experimental results show that our RPM-Net achieves state-of-the-art performance compared to existing non-deep learning and recent deep learning methods. Our source code is available at the project website https://github.com/yewzijian/RPMNet .

For agents

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

get_harvested_code_for_paper("2003.13479")
get_code_for_paper("2003.13479")
have("2003.13479")

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

curl https://syntology.ai/api/ran/2003.13479.json

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

[![Syntology run record](https://syntology.ai/api/ran/2003.13479.svg)](https://syntology.ai/paper/2003.13479)

Connect an agent — have() is free.