SYNTOLOGY HomeExplorerAtlasCodeMethodologyAboutDevelopersFeedPricing
Paper · 2202.03377 · ICML · 2022

Benchmarking and Analyzing Point Cloud Classification under Corruptions

Jiawei Ren, Ziwei Liu, Liang Pan

arXiv · PDF · Open in the Atlas

Code that ran

11 of 13 samples (functions and classes) lifted from this paper's linked repositories executed on a synthesized input in Syntology's jail (0 of them constructed an object rather than computing a result; 0 honoured the contract we drafted, 0 violated it and 5 have no contract verdict; 6 failures of our instrument; 2 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/2202.03377.json · badge: /api/ran/2202.03377.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
jiawei-ren/modelnetc canonical (classifier label)
classifier basis not recorded on the link
11 ran · 2 no recorded run
FunctionStatusWhere it lives
corrupt_jitter Ran
ran on a synthesized input
jiawei-ren/modelnetc/build_modelnetc/corrupt_utils.py
code served (permissive licence) · get_code("b5dd3f043c271b69")
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

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

# Create a small point cloud: N=2 points, C=3 coordinates (x, y, z)
pointcloud = np.random.randn(2, 3).astype(np.float32)

# Define level parameter (severity level, integer)
level = 1

# Construct ARGS dict with all parameters
ARGS = {
    "pointcloud": pointcloud,
    "level": level
}
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
corrupt_rotate Ran
ran on a synthesized input
jiawei-ren/modelnetc/build_modelnetc/corrupt_utils.py
code served (permissive licence) · get_code("a260494677ed35d8")
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
import math

np.random.seed(0)

# Create a small point cloud: 4 points with 3 coordinates each
pointcloud = np.random.randn(4, 3).astype(np.float32)

# Set level to a valid severity level (0-4 typically)
level = 2

ARGS = {
    "pointcloud": pointcloud,
    "level": level,
}
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
corrupt_scale Ran
ran on a synthesized input
jiawei-ren/modelnetc/build_modelnetc/corrupt_utils.py
code served (permissive licence) · get_code("8f0a929253e9f007")
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

np.random.seed(0)

# Create a small point cloud: shape (5, 3) representing 5 points with 3 coordinates (x, y, z)
pointcloud = np.random.randn(5, 3).astype('float32')

# Level should be an integer from 0-4 (indexing into the list in corrupt_scale)
level = 0

ARGS = {
    'pointcloud': pointcloud,
    'level': level,
}
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
cut_points Ran
ran on a synthesized input
jiawei-ren/modelnetc/GDANet/rsmix_provider.py
code served (permissive licence) · get_code("d26265da343f8191")
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

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

# Create small test data
# data_batch: [B, N, C] where B=batch, N=num_points, C=coordinates (or coordinates+normals)
# Using B=2, N=10, C=6 (3 for xyz + 3 for normals)
B = 2
N = 10
C = 6

data_batch = np.random.randn(B, N, C).astype(np.float32)

# idx: [B, S] where S=1 (one query point per batch)
# Contains indices between 0 and N-1
idx = np.array([[3], [5]], dtype=np.int32)

# radius: local region radius (scalar float)
radius = 2.0

# nsample: max sample number in local region (scalar int)
nsample = 5

ARGS = {
    "data_batch": data_batch,
    "idx": idx,
    "radius": radius,
    "nsample": nsample
}
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
cut_points_knn Ran
ran on a synthesized input
jiawei-ren/modelnetc/GDANet/rsmix_provider.py
code served (permissive licence) · get_code("f3bf3d79146b55df")
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

# Create small test data
B = 2  # batch size
N = 8  # number of points
C = 3  # coordinates (x, y, z)
S = 1  # query size (one index per batch)

# Create random point cloud data (B, N, C)
np.random.seed(0)
data_batch = np.random.randn(B, N, C).astype(np.float32)

# Create indices for query (B, S) - values between 0 and N-1
idx = np.array([[2], [5]], dtype=np.int64)

# Other parameters
radius = 1.0
nsample = 4
k = 4

ARGS = {
    "data_batch": data_batch,
    "idx": idx,
    "radius": radius,
    "nsample": nsample,
    "k": k
}
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
get_graph_feature Ran
ran; the contract check could not run (RAISES, DEP_MISSING or TIMEOUT)
jiawei-ren/modelnetc/PointWOLF/model.py
code served (permissive licence) · get_code("9a8756778dd8234a")
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; no output shape is stated. The image it ran in was not recorded. The code mentions cuda, and the harness aliases cuda to the CPU for every sample; whether this run needed that is 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)

# Construct a small tensor for x
# Shape: (batch_size, num_dims, num_points)
# batch_size=2, num_dims=4, num_points=3
x = torch.randn(2, 4, 3)

# k parameter for knn
k = 2

# idx parameter (optional, can be None to let function compute it)
idx = None

ARGS = {
    "x": x,
    "k": k,
    "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 4, dtype family float — no output shape is stated: observe_signature.py could not observe it (call:AssertionError)
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
knn Ran
ran; the contract check could not run (RAISES, DEP_MISSING or TIMEOUT)
jiawei-ren/modelnetc/PointWOLF/model.py
code served (permissive licence) · get_code("cdd0141594039dcb")
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, 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)

# x is used with:
# - transpose(2, 1): expects 3D tensor
# - matmul with transposed version: matrix multiplication
# - sum(x ** 2, dim=1, keepdim=True): sum over dimension 1
# - topk(k=k, dim=-1): get top k indices along last dimension
# 
# Shape should be (batch, features, num_points) based on the transpose operations
# Let's use batch=2, features=4, num_points=3

x = torch.randn(2, 4, 3)

# k is the number of nearest neighbors to find
# Should be less than num_points (3), so k=2
k = 2

ARGS = {
    "x": x,
    "k": k,
}
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, 2) — shape (2, 3, 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
load_h5 Ran
ran; our contract draft was wrong, not the code
jiawei-ren/modelnetc/modelnetc_utils/modelnetc_utils/dataset.py
pointer only (licence: Apache-2.0); the stored copy get_code serves, from yossilevii100/critical_points2, is pointer only (licence: NONE) · get_code("1bfbbe6bae6f8e29")
How it ran

The body of this sample is withheld: this paper's link allows the body inline (Apache-2.0), but get_code serves the stored copy, first harvested from yossilevii100/critical_points2, 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.

rotate_point_cloud Ran
ran; the contract check could not run (RAISES, DEP_MISSING or TIMEOUT)
jiawei-ren/modelnetc/GDANet/provider.py
code served (permissive licence) · get_code("adb85367d5bf360a")
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 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 numpy as np

np.random.seed(0)

# Create a small batch of point clouds
# BxNx3 array where B=2 (batch size), N=4 (number of points), 3 (x,y,z coordinates)
batch_data = np.random.randn(2, 4, 3).astype(np.float32)

ARGS = {
    "batch_data": batch_data
}
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 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
shuffle_data Ran
ran; our contract draft was wrong, not the code
jiawei-ren/modelnetc/GDANet/provider.py
pointer only (licence: Apache-2.0); the stored copy get_code serves, from Levishery/Flywire-Neuron-Tracing, is pointer only (licence: NONE) · get_code("06353aadebc1724b")
How it ran

The body of this sample is withheld: this paper's link allows the body inline (Apache-2.0), but get_code serves the stored copy, first harvested from Levishery/Flywire-Neuron-Tracing, 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.

shuffle_points Ran
ran; the contract check could not run (RAISES, DEP_MISSING or TIMEOUT)
jiawei-ren/modelnetc/GDANet/provider.py
code served (permissive licence) · get_code("a97f76b9c0811c93")
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 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 numpy as np

# Create a small batch of point cloud data
# BxNxC format: B=batch size, N=number of points, C=channels (coordinates)
batch_size = 2
num_points = 4
channels = 3

batch_data = np.random.randn(batch_size, num_points, channels).astype(np.float32)

ARGS = {
    "batch_data": batch_data
}
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 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
knn_points Not yet run
unverified (harvested, no recorded run)
jiawei-ren/modelnetc/GDANet/rsmix_provider.py
code served (permissive licence) · get_code("a7032e8191925cc0")
How it ran

no invocation recorded.

load_data Not yet run
unverified (harvested, no recorded run)
jiawei-ren/modelnetc/build_modelnetc/corrupt.py
code served (permissive licence) · get_code("6002e0aaef533772")
How it ran

no invocation recorded.

Also present in other papers' code

Of 13 samples lifted from this paper's linked repositories, 13 carry a function identity and 0 do not. Of the 13 distinct functions with an identity, 6 are also present in other papers whose code we hold (5 with at least one paper dated earlier, 0 the same month, 6 later) and 7 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
knn 52 other papers
15 dated earlier, 37 later · 12 only as an assembly brick
1801.07829 (2018-01)
Dynamic Graph CNN for Learning on Point Clouds
1905.06292 (2019-05)
3D Point Cloud Generative Adversarial Network Based on Tree Structured Graph Convolutions
2005.06734 (2020-05)
Dense-Resolution Network for Point Cloud Classification and Segmentation
and 49 more not named here
shuffle_data 17 other papers
12 dated earlier, 4 later, 1 same year, order not determinable
1803.05827 (2018-03)
Local Spectral Graph Convolution for Point Set Feature Learning
1803.11527 (2018-03)
SpiderCNN: Deep Learning on Point Sets with Parameterized Convolutional Filters
1811.07246 (2018-11)
PointConv: Deep Convolutional Networks on 3D Point Clouds
and 14 more not named here
get_graph_feature 13 other papers
4 dated earlier, 9 later · 2 only as an assembly brick
1801.07829 (2018-01)
Dynamic Graph CNN for Learning on Point Clouds
2109.00179 (2021-09)
Spatio-temporal Self-Supervised Representation Learning for 3D Point Clouds
2111.04426 (2021-11)
3D Siamese Voxel-to-BEV Tracker for Sparse Point Clouds
and 10 more not named here
shuffle_points 11 other papers
6 dated earlier, 4 later, 1 same year, order not determinable
1811.07246 (2018-11)
PointConv: Deep Convolutional Networks on 3D Point Clouds
1812.11647 (2018-12)
Path-Invariant Map Networks
2006.14865 (2020-06)
RPM-Net: Recurrent Prediction of Motion and Parts from Point Cloud
and 8 more not named here
rotate_point_cloud 7 other papers
5 dated earlier, 1 later, 1 same year, order not determinable
1803.11527 (2018-03)
SpiderCNN: Deep Learning on Point Sets with Parameterized Convolutional Filters
1904.10014 (2019-04)
Linked Dynamic Graph CNN: Learning on Point Cloud via Linking Hierarchical Features
1908.04616 (2019-08)
Revisiting Point Cloud Classification: A New Benchmark Dataset and Classification Model on Real-World Data
and 4 more not named here
load_h5 1 other paper
1 later
2308.05525 (2023-08)
Robustifying Point Cloud Networks by Refocusing

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

3D perception, especially point cloud classification, has achieved substantial progress. However, in real-world deployment, point cloud corruptions are inevitable due to the scene complexity, sensor inaccuracy, and processing imprecision. In this work, we aim to rigorously benchmark and analyze point cloud classification under corruptions. To conduct a systematic investigation, we first provide a taxonomy of common 3D corruptions and identify the atomic corruptions. Then, we perform a comprehensive evaluation on a wide range of representative point cloud models to understand their robustness and generalizability. Our benchmark results show that although point cloud classification performance improves over time, the state-of-the-art methods are on the verge of being less robust. Based on the obtained observations, we propose several effective techniques to enhance point cloud classifier robustness. We hope our comprehensive benchmark, in-depth analysis, and proposed techniques could spark future research in robust 3D perception. Code is available at https://github. com/jiawei-ren/modelnetc.

For agents

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

get_harvested_code_for_paper("2202.03377")
get_code_for_paper("2202.03377")
have("2202.03377")

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

curl https://syntology.ai/api/ran/2202.03377.json

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

[![Syntology run record](https://syntology.ai/api/ran/2202.03377.svg)](https://syntology.ai/paper/2202.03377)

Connect an agent — have() is free.