SYNTOLOGY HomeExplorerAtlasCodeMethodologyAboutDevelopersFeedPricing
Paper · 2103.10814 · CVPR · 2021

Skeleton Merger: an Unsupervised Aligned Keypoint Detector

Cewu Lu, Yang You, Ruoxi Shi, Zhengrong Xue

arXiv · PDF · Open in the Atlas

Code that ran

7 of 9 samples (functions and classes) lifted from this paper's linked repositories executed on a synthesized input in Syntology's jail (2 of them constructed an object rather than computing a result; 0 honoured the contract we drafted, 0 violated it and 3 have no contract verdict; 4 failures of our instrument; 2 no recorded run), measured 2026-11-22; 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/2103.10814.json · badge: /api/ran/2103.10814.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
eliphatfs/SkeletonMerger marked official in Papers with Code
Papers with Code archive
7 ran · 2 no recorded run
FunctionStatusWhere it lives
Head Ran
ran on a synthesized input
eliphatfs/SkeletonMerger/merger/merger_net.py
code served (permissive licence) · get_code("3c1cee99418ac507")
How it ran

Ran in Syntology's sandbox on this synthesized input (drafted by a language model, class construct-then-call prompt, model not recorded) 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 torch
import numpy as np

INIT_ARGS = {}

ARGS = {
    "KPA": torch.randn(2, 4, 3),
    "KPB": torch.randn(2, 4, 3)
}
Setup block drafted by
a language model, class construct-then-call prompt, model not recorded
Imports
numpy, torch — import names as written, not package names to install; a name the repository itself provides is not a package
Returned
type tuple — 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: torch and numpy seeded to 0 by the harness immediately before the class was constructed from INIT_ARGS (fixture_check.py)
PointNetFeaturePropagation Ran
ran on a synthesized input · constructed an object; no result was computed
eliphatfs/SkeletonMerger/merger/merger_net.py
code served (permissive licence) · get_code("bcb0d966a2004d18")
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 an instance of its own class, PointNetFeaturePropagation: the run constructed the object and did not call it, so no result was computed. 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)

# Create a small PointNetFeaturePropagation instance
in_channel = 5
mlp = [16, 8]

model = nn.Module()

# Create example tensors with distinct sizes
# Batch size: 2, spatial dimensions vary to distinguish them
B = 2
N = 6  # number of points in xyz1
S = 4  # number of points in xyz2
C = 3  # coordinate dimension
D = 5  # feature dimension for points2

# xyz1: [B, C, N] - upsampling targets
xyz1 = torch.randn(B, C, N)

# xyz2: [B, C, S] - sampled/downsampled points
xyz2 = torch.randn(B, C, S)

# points1: [B, D, N] - features at xyz1
points1 = torch.randn(B, D, N)

# points2: [B, D, S] - features at xyz2
points2 = torch.randn(B, D, S)

ARGS = {
    'in_channel': in_channel,
    'mlp': mlp,
}

# For forward pass, we need to instantiate and call it
model_instance = nn.Module()
fp_module = nn.Sequential()

FORWARD_ARGS = {
    'xyz1': xyz1,
    'xyz2': xyz2,
    'points1': points1,
    'points2': points2,
}
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 PointNetFeaturePropagation — the run constructed an instance of its own class and did not call it; no shape is stated — the instrument coerces any output to an array; for a PointNetFeaturePropagation 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
PointNetSetAbstractionMsg Ran
ran on a synthesized input · constructed an object; no result was computed
eliphatfs/SkeletonMerger/merger/merger_net.py
code served (permissive licence) · get_code("aa80a5efbd478b8a")
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 an instance of its own class, PointNetSetAbstractionMsg: the run constructed the object and did not call it, so no result was computed. 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 = {
    "npoint": 8,
    "radius_list": [0.1, 0.2],
    "nsample_list": [16, 32],
    "in_channel": 3,
    "mlp_list": [
        [64, 128],
        [64, 128]
    ]
}
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 PointNetSetAbstractionMsg — the run constructed an instance of its own class and did not call it; no shape is stated — the instrument coerces any output to an array; for a PointNetSetAbstractionMsg 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
farthest_point_sample Ran
ran; the contract check could not run (RAISES, DEP_MISSING or TIMEOUT)
eliphatfs/SkeletonMerger/merger/merger_net.py
code served (permissive licence) · get_code("8016bf9f4fcc1ce4")
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 int, shape (2, 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)

# Construct arguments for farthest_point_sample
# xyz: pointcloud data of shape [B, N, 3]
# npoint: number of samples to select

B = 2  # batch size
N = 4  # number of points
npoint = 2  # number of points to sample

# Create random point cloud data [B, N, 3]
xyz = torch.randn(B, N, 3)

ARGS = {
    "xyz": xyz,
    "npoint": npoint,
}
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 int, shape (2, 2) — shape (2, 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
index_points Ran
ran; the contract check could not run (RAISES, DEP_MISSING or TIMEOUT)
eliphatfs/SkeletonMerger/merger/merger_net.py
code served (permissive licence) · get_code("573cb024ea47999f")
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) and returned Tensor, ndim 3, dtype family float, shape (2, 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 torch

torch.manual_seed(0)

# Create sample tensors
# points: [B, N, C] - batch of point clouds
# B=2 (batch size), N=4 (number of points), C=3 (coordinates)
points = torch.randn(2, 4, 3)

# idx: [B, S] - indices to sample from points
# B=2 (batch size), S=2 (number of samples to select)
idx = torch.tensor([[0, 2], [1, 3]], dtype=torch.long)

ARGS = {
    "points": points,
    "idx": idx,
}
Setup block drafted by
a language model, Claude Haiku 4.5, prompt python_fixture_v1
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, 2, 3) — shape (2, 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
query_ball_point Ran
ran; the contract check could not run (RAISES, DEP_MISSING or TIMEOUT)
eliphatfs/SkeletonMerger/merger/merger_net.py
code served (permissive licence) · get_code("45a0ca2d012bb9fe")
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) and returned Tensor, ndim 3, dtype family int, 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

torch.manual_seed(0)

# Parameters for query_ball_point
radius = 0.5
nsample = 3
xyz = torch.randn(2, 4, 3)  # [B=2, N=4, C=3]
new_xyz = torch.randn(2, 3, 3)  # [B=2, S=3, C=3]

ARGS = {
    "radius": radius,
    "nsample": nsample,
    "xyz": xyz,
    "new_xyz": new_xyz,
}
Setup block drafted by
a language model, Claude Haiku 4.5, prompt python_fixture_v1
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 int, 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
square_distance Ran
ran; the contract check could not run (RAISES, DEP_MISSING or TIMEOUT)
eliphatfs/SkeletonMerger/merger/merger_net.py
code served (permissive licence) · get_code("580dee6761649b29")
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) and returned Tensor, ndim 3, dtype family float, shape (2, 4, 5) 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 source points: [B, N, C]
# B=2 (batch), N=4 (num source points), C=3 (coordinates)
src = torch.randn(2, 4, 3)

# Create destination points: [B, M, C]
# B=2 (batch), M=5 (num destination points), C=3 (coordinates)
dst = torch.randn(2, 5, 3)

ARGS = {
    "src": src,
    "dst": dst,
}
Setup block drafted by
a language model, Claude Haiku 4.5, prompt python_fixture_v1
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, 5) — shape (2, 4, 5) 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
Net Not yet run
unverified (harvested, no recorded run)
eliphatfs/SkeletonMerger/merger/merger_net.py
code served (permissive licence) · get_code("fe58e66f9a261a66")
How it ran

no invocation recorded.

PBlock Not yet run
unverified (harvested, no recorded run)
eliphatfs/SkeletonMerger/merger/merger_net.py
code served (permissive licence) · get_code("e63c2e0accc60734")
How it ran

no invocation 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, 6 are also present in other papers whose code we hold (4 with at least one paper dated earlier, 1 the same month, 6 later) and 3 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
square_distance
brick of this paper's assembly: helper
75 other papers
5 dated earlier, 1 the same month, 69 later · 9 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
2003.00410 (2020-03)
PF-Net: Point Fractal Network for 3D Point Cloud Completion
and 72 more not named here
index_points
brick of this paper's assembly: helper
42 other papers
6 dated earlier, 36 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
2003.13479 (2020-03)
RPM-Net: Robust Point Matching using Learned Features
and 39 more not named here
farthest_point_sample
brick of this paper's assembly: helper
9 other papers
3 dated earlier, 6 later · 4 only as an assembly brick
1612.00593 (2016-12)
PointNet: Deep Learning on Point Sets for 3D Classification and Segmentation
1706.02413 (2017-06)
PointNet++: Deep Hierarchical Feature Learning on Point Sets in a Metric Space
2004.10904 (2020-04)
Through the Looking Glass: Neural 3D Reconstruction of Transparent Shapes
and 6 more not named here
query_ball_point
brick of this paper's assembly: helper
7 other papers
1 dated earlier, 6 later · 2 only as an assembly brick
1706.02413 (2017-06)
PointNet++: Deep Hierarchical Feature Learning on Point Sets in a Metric Space
2206.04670 (2022-06)
PointNeXt: Revisiting PointNet++ with Improved Training and Scaling Strategies
2303.10437 (2023-03)
Grounding 3D Object Affordance from 2D Interactions in Images
and 4 more not named here
PointNetFeaturePropagation
brick of this paper's assembly: helper
3 other papers
3 later
2303.10437 (2023-03)
Grounding 3D Object Affordance from 2D Interactions in Images
2305.16963 (2023-05)
Semantic segmentation of sparse irregular point clouds for leaf/wood discrimination
2306.00777 (2023-06)
Object pop-up: Can we infer 3D objects and their poses from human interactions alone?
PointNetSetAbstractionMsg
brick of this paper's assembly: helper
2 other papers
2 later
2303.10437 (2023-03)
Grounding 3D Object Affordance from 2D Interactions in Images
2306.00777 (2023-06)
Object pop-up: Can we infer 3D objects and their poses from human interactions alone?

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

Detecting aligned 3D keypoints is essential under many scenarios such as object tracking, shape retrieval and robotics. However, it is generally hard to prepare a highquality dataset for all types of objects due to the ambiguity of keypoint itself. Meanwhile, current unsupervised detectors are unable to generate aligned keypoints with good coverage. In this paper, we propose an unsupervised aligned keypoint detector, Skeleton Merger, which utilizes skeletons to reconstruct objects. It is based on an Autoencoder architecture. The encoder proposes keypoints and predicts activation strengths of edges between keypoints. The decoder performs uniform sampling on the skeleton and refines it into small point clouds with pointwise offsets. Then the activation strengths are applied and the sub-clouds are merged. Composite Chamfer Distance (CCD) is proposed as a distance between the input point cloud and the reconstruction composed of sub-clouds masked by activation strengths. We demonstrate that Skeleton Merger is capable of detecting semantically-rich salient keypoints with good alignment, and shows comparable performance to supervised methods on the KeypointNet dataset. It is also shown that the detector is robust to noise and subsampling. Our code is available at https://github.com/eliphatfs/SkeletonMerger.

For agents

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

get_harvested_code_for_paper("2103.10814")
get_code_for_paper("2103.10814")
have("2103.10814")

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

curl https://syntology.ai/api/ran/2103.10814.json

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

[![Syntology run record](https://syntology.ai/api/ran/2103.10814.svg)](https://syntology.ai/paper/2103.10814)

Connect an agent — have() is free.