SYNTOLOGY HomeExplorerAtlasCodeMethodologyAboutDevelopersFeedPricing
Paper · 2403.01316 · CVPR · 2024

TUMTraf V2X Cooperative Perception Dataset

Rui Song, Alois Knoll, Walter Zimmer, Gerhard Wardana, Suren Sritharan, Id Zhou

arXiv · PDF · Open in the Atlas

Code that ran

6 of 6 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 0 have no contract verdict; 6 failures of our instrument; 0 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/2403.01316.json · badge: /api/ran/2403.01316.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
tum-traffic-dataset/tum-traffic-dataset-dev-kit canonical (classifier label)
classifier basis: classifier_2025
3 ran · 0 no recorded run
tum-traffic-dataset/coopdet3d canonical (classifier label)
classifier basis: classifier_2025
2 ran · 0 no recorded run
walzimmer/3d-bat canonical (classifier label)
classifier basis: census
1 ran · 0 no recorded run
FunctionStatusWhere it lives
detections_to_dict Ran
ran; our contract draft was wrong, not the code
tum-traffic-dataset/tum-traffic-dataset-dev-kit/src/utils/detection.py
code served (permissive licence) · get_code("e1192fe79d2a9fa8")
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 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
from dataclasses import dataclass, field
from typing import Tuple, Optional, List, Any
import math

# Reset the global counter
next_detection_id = 0

@dataclass
class Detection:
    """Representation of a single detected road user."""
    location: np.ndarray
    dimensions: Tuple[float, float, float]
    yaw: float
    category: str
    bottom_contour: Optional[np.ndarray] = None
    yaw_history: List[float] = field(default_factory=list)
    pos_history: List[np.ndarray] = field(default_factory=list)
    bbox_2d: Optional[np.ndarray] = None
    id: int = -1
    sensor_id: str = ''
    uuid: str = ''
    velocity: Optional[np.ndarray] = None
    speed: float = None
    original_detected_object_ros_msg: Optional[Any] = None
    color: Optional[str] = None
    num_lidar_points: int = 0
    score: float = -1.0
    yaw_base: float = 0.0
    occlusion_level: str = None
    overlap: bool = None
    has_trailer: bool = None
    has_rider: bool = None
    has_flashing_lights: bool = None
    number_of_trailers: int = None
    is_electric: bool = None
    sub_type: str = None
    existence_probability: float = None
    yaw_rate: float = None
    box3d_projected = None

    def __post_init__(self):
        global next_detection_id
        if self.id == -1:
            self.id = next_detection_id
        next_detection_id += 1

def detections_to_dict(detection_list: List[Detection]):
    """
    Convert list of detected detections to the format
    expected by the devkit evaluation module.
    """
    names = []
    boxes = []
    scores = []
    for v in detection_list:
        names.append(v.category.capitalize())
        values = list(v.location.flatten()) + list(v.dimensions)
        values.append(v.yaw)
        boxes.append(np.array(values))
        scores.append(v.score)
    return {'name': np.array(names), 'boxes_3d': np.array(boxes), 'score': np.array(scores)}


# Create sample Detection objects
detection1 = Detection(
    location=np.array([1.0, 2.0, 3.0]),
    dimensions=(4.5, 2.0, 1.8),
    yaw=0.5,
    category='car',
    score=0.95
)

detection2 = Detection(
    location=np.array([5.0, 6.0, 7.0]),
    dimensions=(3.5, 1.8, 1.5),
    yaw=1.2,
    category='pedestrian',
    score=0.87
)

ARGS = {
    'detection_list': [detection1, detection2]
}
Setup block drafted by
a language model, Claude Haiku 4.5, prompt python_fixture_v1
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
filter_box_in_lidar_cs Ran
ran; our contract draft was wrong, not the code
tum-traffic-dataset/coopdet3d/mmdet3d/datasets/tumtraf_v2x_dataset.py
code served (permissive licence) · get_code("2b0ad028e517b7ad")
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 list, ndim 1, 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

# Create sample boxes with required structure
boxes = [
    {
        'center': np.array([5.0, 3.0, 1.5]),  # 3D center position
        'label': 0,  # class label index
    },
    {
        'center': np.array([15.0, 8.0, 2.0]),  # Another box farther away
        'label': 1,
    },
    {
        'center': np.array([2.0, 1.0, 0.5]),  # Close box
        'label': 0,
    },
]

# Create class names list
classes = ['car', 'pedestrian']

# Create evaluation config with class range mapping
eval_configs = {
    'class_range': {
        'car': 10.0,  # Detection range for car class
        'pedestrian': 5.0,  # Detection range for pedestrian class
    }
}

ARGS = {
    'boxes': boxes,
    'classes': classes,
    'eval_configs': eval_configs,
}
Setup block drafted by
a language model, Claude Haiku 4.5, prompt python_fixture_v1
Imports
numpy — import names as written, not package names to install; a name the repository itself provides is not a package
Returned
type list, ndim 1, dtype family float — a list output has no array shape of its own; none is stated
dtype family
int means an integer dtype; float means any other dtype the instrument saw, bool and complex included (fixture_check.py records only these two families)
Environment
image not recorded; re-verify against syntology-harvest-sandbox:v5-deps; not recorded for this sample: the harness installs a cuda-to-cpu shim (cuda_shim.py) before every sample and no run records whether a sample needed it; seed: none by the harness; any seed is in the setup block
get_corners Ran
ran; the contract check could not run (RAISES, DEP_MISSING or TIMEOUT)
tum-traffic-dataset/tum-traffic-dataset-dev-kit/src/utils/detection.py
code served (permissive licence) · get_code("caff23ff5551c710")
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 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 arguments for get_corners function
yaw = 0.5  # float: rotation angle in radians
width = 2.0  # float: width of the bounding box
length = 3.0  # float: length of the bounding box
position = np.array([1.0, 2.0, 0.0])  # np.ndarray: 3D position

ARGS = {
    "yaw": yaw,
    "width": width,
    "length": length,
    "position": position
}
Setup block drafted by
a language model, Claude Haiku 4.5, prompt python_fixture_v1
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
output_to_box_dict Ran
ran; our contract draft was wrong, not the code
tum-traffic-dataset/coopdet3d/mmdet3d/datasets/tumtraf_v2x_dataset.py
code served (permissive licence) · get_code("735a536152d7d9d1")
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 list, ndim 1, dtype family float; no output shape is stated. The image it ran in was not recorded. If your copy returns the same type and ndim on these inputs, it matched what we observed; nothing here checks that. It does NOT mean the output is correct: nothing has compared it to the paper.

import torch
import numpy as np

torch.manual_seed(0)

class BaseInstance3DBoxes:
    """Stand-in class for 3D bounding boxes."""
    def __init__(self, tensor):
        self.tensor = tensor
        self.gravity_center = tensor[:, :3]
        self.dims = tensor[:, 3:6]
        self.yaw = tensor[:, 6:7].squeeze(-1)
    
    def __len__(self):
        return len(self.tensor)

detection = {
    'boxes_3d': BaseInstance3DBoxes(
        torch.tensor([
            [1.0, 2.0, 3.0, 4.0, 5.0, 6.0, 0.5, 0.1, 0.2],
            [7.0, 8.0, 9.0, 10.0, 11.0, 12.0, 1.5, 0.3, 0.4]
        ], dtype=torch.float32)
    ),
    'scores_3d': torch.tensor([0.95, 0.87], dtype=torch.float32),
    'labels_3d': torch.tensor([0, 1], dtype=torch.float32)
}

ARGS = {
    'detection': detection
}
Setup block drafted by
a language model, Claude Haiku 4.5, prompt python_fixture_v1
Imports
numpy, torch — import names as written, not package names to install; a name the repository itself provides is not a package
Returned
type list, ndim 1, dtype family float — a list output has no array shape of its own; none is stated
dtype family
int means an integer dtype; float means any other dtype the instrument saw, bool and complex included (fixture_check.py records only these two families)
Environment
image not recorded; re-verify against syntology-harvest-sandbox:v5-deps; not recorded for this sample: the harness installs a cuda-to-cpu shim (cuda_shim.py) before every sample and no run records whether a sample needed it; seed: none by the harness; any seed is in the setup block
sample_one_obj Ran
ran; the contract check could not run (RAISES, DEP_MISSING or TIMEOUT)
walzimmer/3d-bat/src/server/pre_annotate.py
pointer only (licence: NOASSERTION) · get_code("a5e80a6897ac1f7e")
How it ran

The body of this sample is withheld: GitHub could not classify the licence of this paper's repository, walzimmer/3d-bat (NOASSERTION). 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.

save_to_openlabel Ran
ran; our contract draft was wrong, not the code
tum-traffic-dataset/tum-traffic-dataset-dev-kit/src/utils/detection.py
code served (permissive licence) · get_code("232caac9aac3c2b2")
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 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
from pathlib import Path
from dataclasses import dataclass, field
from typing import List, Optional, Tuple, Any

# Define the Detection class (copied from the function signature)
@dataclass
class Detection:
    """Representation of a single detected road user."""
    location: np.ndarray
    dimensions: Tuple[float, float, float]
    yaw: float
    category: str
    bottom_contour: Optional[np.ndarray] = None
    yaw_history: List[float] = field(default_factory=list)
    pos_history: List[np.ndarray] = field(default_factory=list)
    bbox_2d: Optional[np.ndarray] = None
    id: int = -1
    sensor_id: str = ''
    uuid: str = ''
    velocity: Optional[np.ndarray] = None
    speed: float = None
    original_detected_object_ros_msg: Optional[Any] = None
    color: Optional[str] = None
    num_lidar_points: int = 0
    score: float = -1.0
    yaw_base: float = 0.0
    occlusion_level: str = None
    overlap: bool = None
    has_trailer: bool = None
    has_rider: bool = None
    has_flashing_lights: bool = None
    number_of_trailers: int = None
    is_electric: bool = None
    sub_type: str = None
    existence_probability: float = None
    yaw_rate: float = None
    box3d_projected = None

    def __post_init__(self):
        pass

# Create sample detections
detection1 = Detection(
    location=np.array([1.0, 2.0, 0.5]),
    dimensions=(2.0, 1.5, 1.8),
    yaw=0.5,
    category="car",
    velocity=np.array([0.1, 0.2, 0.0]),
    bbox_2d=np.array([100, 150, 200, 250])
)

detection2 = Detection(
    location=np.array([5.0, 3.0, 0.5]),
    dimensions=(1.8, 1.2, 1.6),
    yaw=1.2,
    category="pedestrian",
    velocity=np.array([0.05, 0.1, 0.0]),
    bbox_2d=np.array([300, 200, 350, 400])
)

detection_list = [detection1, detection2]

# Create coordinate systems dict
coordinate_systems = {
    "sensor": {
        "type": "sensor",
        "parent": "world"
    },
    "world": {
        "type": "world"
    }
}

# Create frame properties dict
frame_properties = {
    "timestamp": 1234567890
}

# Create streams dict
streams = {
    "camera": {
        "type": "camera"
    }
}

ARGS = {
    "detection_list": detection_list,
    "filename": "detections.json",
    "output_folder_path": Path("/tmp"),
    "coordinate_systems": coordinate_systems,
    "frame_properties": frame_properties,
    "frame_id": "0",
    "streams": streams
}
Setup block drafted by
a language model, Claude Haiku 4.5, prompt python_fixture_v1
Imports
numpy, scipy — 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

Also present in other papers' code

Of 6 samples lifted from this paper's linked repositories, 6 carry a function identity and 0 do not. Of the 6 distinct functions with an identity, 2 are also present in other papers whose code we hold (2 with at least one paper dated earlier, 0 the same month, 0 later) and 4 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
filter_box_in_lidar_cs 1 other paper
1 dated earlier
2203.06319 (2022-03)
PillarGrid: Deep Learning-based Cooperative Perception for 3D Object Detection from Onboard-Roadside LiDAR
output_to_box_dict 1 other paper
1 dated earlier
2203.06319 (2022-03)
PillarGrid: Deep Learning-based Cooperative Perception for 3D Object Detection from Onboard-Roadside LiDAR

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

Registered Point Cloud Roadside cameras Roadside LiDAR Vehicle camera Roadside camera Roadside camera Roadside camera Vehicle LiDAR Figure 1. Visualization of 3D box labels and tracks in our TUMTraf V2X Cooperative Perception Dataset. The top part shows the labels projected into the four camera images. The part below shows a point cloud from two LiDARs with 3D box labels of the same scene.

For agents

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

get_harvested_code_for_paper("2403.01316")
get_code_for_paper("2403.01316")
have("2403.01316")

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

curl https://syntology.ai/api/ran/2403.01316.json

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

[![Syntology run record](https://syntology.ai/api/ran/2403.01316.svg)](https://syntology.ai/paper/2403.01316)

Connect an agent — have() is free.