SYNTOLOGY HomeExplorerAtlasCodeMethodologyAboutDevelopersFeedPricing
Paper · 2303.13174 · CVPR · 2023

3D-POP -An automated annotation approach to facilitate markerless 2D-3D tracking of freely moving birds with marker-based motion capture

Hemal Naik, Alex Hoi, Hang Chan, Junran Yang, Mathilde Delacoux, Iain Couzin, Fumihiro Kano, Máté Nagy

arXiv · PDF · Open in the Atlas

Code that ran

13 of 15 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 13 have no contract verdict; 0 failures of our instrument; 2 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/2303.13174.json · badge: /api/ran/2303.13174.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
alexhang212/Dataset-3DPOP canonical (classifier label)
classifier basis not recorded on the link
13 ran · 2 no recorded run
FunctionStatusWhere it lives
Extract3DPoints Ran
ran on a synthesized input
alexhang212/Dataset-3DPOP/POP3D_AP/ApplicationExamples/ComputePlanes.py
code served (permissive licence) · get_code("0c48eb5f51fc8c81")
How it ran

Ran in Syntology's sandbox on this synthesized input (drafted by a language model, Claude Haiku 4.5, prompt python_fixture_v1, through the batch API) and returned dict. The image it ran in was not recorded. If your copy returns the same type on these inputs, it matched what we observed; nothing here checks that. It does NOT mean the output is correct: nothing has compared it to the paper.

import numpy as np

class ViconObject:
    def __init__(self, name):
        self.name = name

# Test case 1: head object (ends with 'hd')
viconObject_head = ViconObject('bird_hd')
rowData_head = {
    'bird_hd_beak_x': 1.0,
    'bird_hd_beak_y': 2.0,
    'bird_hd_beak_z': 3.0,
    'bird_hd_leftEye_x': 1.5,
    'bird_hd_leftEye_y': 2.5,
    'bird_hd_leftEye_z': 3.5,
    'bird_hd_rightEye_x': 0.5,
    'bird_hd_rightEye_y': 1.5,
    'bird_hd_rightEye_z': 2.5,
}

ARGS = {
    'viconObject': viconObject_head,
    'rowData': rowData_head
}
Setup block drafted by
a language model, Claude Haiku 4.5, prompt python_fixture_v1, through the batch API
Imports
numpy — import names as written, not package names to install; a name the repository itself provides is not a package
Returned
type dict — a dict output has no array shape of its own; none is stated — the instrument coerces any output to an array; for a dict output the recorded ndim 0 and dtype family float describe that coercion, not the output, and are not stated
Environment
image not recorded; re-verify against syntology-harvest-sandbox:v5-deps; not recorded for this sample: the harness installs a cuda-to-cpu shim (cuda_shim.py) before every sample and no run records whether a sample needed it; seed: none by the harness; any seed is in the setup block
FindClosestFrame Ran
ran on a synthesized input
alexhang212/Dataset-3DPOP/POP3D_AP/ApplicationExamples/ManualAnnot.py
code served (permissive licence) · get_code("86e2483d1bf5526b")
How it ran

Ran in Syntology's sandbox on this synthesized input (drafted by a language model, Claude Haiku 4.5, prompt python_fixture_v1, through the batch API) and returned float64, ndim 0, dtype family float; no output shape is stated. The image it ran in was not recorded. If your copy returns the same type and ndim on these inputs, it matched what we observed; nothing here checks that. It does NOT mean the output is correct: nothing has compared it to the paper.

import numpy as np

# Construct a SyncArray with two rows:
# First row: Sony frame timestamps
# Second row: Corresponding Vicon frame indices
# We need at least 3 elements to avoid index out of bounds when accessing index+1
SyncArray = np.array([
    [100.0, 200.0, 300.0, 400.0],  # Sony frame timestamps
    [10.0, 20.0, 30.0, 40.0]        # Vicon frame indices
], dtype=np.float64)

# Frame should be between two Sony timestamps to properly interpolate
# Let's choose a frame between 100 and 200
Frame = 150.0

ARGS = {
    "SyncArray": SyncArray,
    "Frame": Frame
}
Setup block drafted by
a language model, Claude Haiku 4.5, prompt python_fixture_v1, through the batch API
Imports
numpy — import names as written, not package names to install; a name the repository itself provides is not a package
Returned
type float64, ndim 0, dtype family float — no output shape is stated: observe_signature.py recorded float ndim 0, which disagrees with the fixture run
dtype family
int means an integer dtype; float means any other dtype the instrument saw, bool and complex included (fixture_check.py records only these two families)
Environment
image not recorded; re-verify against syntology-harvest-sandbox:v5-deps; not recorded for this sample: the harness installs a cuda-to-cpu shim (cuda_shim.py) before every sample and no run records whether a sample needed it; seed: none by the harness; any seed is in the setup block
Get_Dict Ran
ran on a synthesized input
alexhang212/Dataset-3DPOP/POP3D_AP/ApplicationExamples/GapFillHybridExample.py
code served (permissive licence) · get_code("1d8dfecf148cffac")
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, 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 pandas as pd
import numpy as np

# Create a multi-index DataFrame for df (DLC predictions)
# Structure: columns are (bodypart_index, keypoint, coordinate)
index_name = 'DLC_resnet50_mouse_trackingJul23shuffle1_1030000'
keypoints = ['nose', 'left_ear', 'right_ear', 'tail_base']
coordinates = ['x', 'y', 'likelihood']

# Create multi-index columns
arrays = [
    [index_name] * len(keypoints) * len(coordinates),
    keypoints * len(coordinates),
    coordinates * len(keypoints)
]
columns = pd.MultiIndex.from_arrays(arrays)

# Create DataFrame with sample data
data = {}
for kp in keypoints:
    for coord in coordinates:
        data[(index_name, kp, coord)] = [100.5, 200.3, 150.2]

df = pd.DataFrame(data)
df.index = ['frame-0.jpg', 'frame-1.jpg', 'frame-2.jpg']

# Create BBoxdf with bounding box data
bbox_data = {
    '452_0107_BBox_x': [50, 60, 55],
    '452_0107_BBox_y': [40, 45, 42],
    '452_0107_BBox_w': [200, 210, 205],
    '452_0107_BBox_h': [180, 190, 185]
}
BBoxdf = pd.DataFrame(bbox_data, index=[0, 1, 2])

frameNo = 0
imsize = (320, 320)

ARGS = {
    'df': df,
    'BBoxdf': BBoxdf,
    'frameNo': frameNo,
    'imsize': imsize
}
Setup block drafted by
a language model, Claude Haiku 4.5, prompt python_fixture_v1, through the batch API
Imports
numpy, pandas — import names as written, not package names to install; a name the repository itself provides is not a package
Returned
type tuple, ndim 1, dtype family float — a tuple 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
IntroduceGap Ran
ran on a synthesized input
alexhang212/Dataset-3DPOP/POP3D_AP/ApplicationExamples/GapFillHybridExample.py
code served (permissive licence) · get_code("42b706274d0d33ec")
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 DataFrame. 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 pandas as pd
import numpy as np
import random
import math

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

# Create a sample DataFrame that mimics CSV3Ddf
# The function expects:
# - A DataFrame with at least 350+ rows (it slices from index 350)
# - A 'Cam1frame' column (which should NOT be filled with NaN)
# - Other columns that will be filled with NaN during gap introduction
# - Some NaN values that will be dropped

data = {
    'Cam1frame': np.arange(400),
    'feature1': np.random.randn(400),
    'feature2': np.random.randn(400),
    'feature3': np.random.randn(400),
}

CSV3Ddf = pd.DataFrame(data)

# Introduce some NaN values so dropna() has something to work with
CSV3Ddf.loc[100:105, 'feature1'] = np.nan
CSV3Ddf.loc[200:202, 'feature2'] = np.nan

ARGS = {
    'CSV3Ddf': CSV3Ddf
}
Setup block drafted by
a language model, Claude Haiku 4.5, prompt python_fixture_v1, through the batch API
Imports
numpy, pandas — import names as written, not package names to install; a name the repository itself provides is not a package
Returned
type DataFrame — a DataFrame output has no array shape of its own; none is stated — the instrument coerces any output to an array; for a DataFrame output the recorded ndim 2 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
IsPointValid Ran
ran on a synthesized input
alexhang212/Dataset-3DPOP/Util/VisualizeUtil.py
code served (permissive licence) · get_code("7d061d73a51976a7")
How it ran

Ran in Syntology's sandbox on this synthesized input (drafted by a language model, Claude Haiku 4.5, prompt python_fixture_v1, through the batch API) and returned bool, ndim 0, dtype family float, shape () as a later run recorded it. The image it ran in was not recorded. If your copy returns the same type, ndim and shape on these inputs, it matched what we observed; nothing here checks that. It does NOT mean the output is correct: nothing has compared it to the paper.

import torch

torch.manual_seed(0)

# Dim: tuple or list representing image dimensions (width, height)
Dim = (100, 150)

# point: tuple or list representing a point [x, y]
point = [50, 75]

ARGS = {
    "Dim": Dim,
    "point": point
}
Setup block drafted by
a language model, Claude Haiku 4.5, prompt python_fixture_v1, through the batch API
Imports
torch — import names as written, not package names to install; a name the repository itself provides is not a package
Returned
type bool, ndim 0, dtype family float, shape () — shape () as a later run recorded it
dtype family
int means an integer dtype; float means any other dtype the instrument saw, bool and complex included (fixture_check.py records only these two families)
Environment
image not recorded; re-verify against syntology-harvest-sandbox:v5-deps; not recorded for this sample: the harness installs a cuda-to-cpu shim (cuda_shim.py) before every sample and no run records whether a sample needed it; seed: none by the harness; any seed is in the setup block
countGap Ran
ran on a synthesized input
alexhang212/Dataset-3DPOP/POP3D_AP/ApplicationExamples/GapFillHybridExample.py
code served (permissive licence) · get_code("eaf20efbb95a0c37")
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 list, ndim 1, dtype family int; 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

def countGap(a):
    """Count nan gaps"""
    tmp = 0
    GapList = []
    for i in range(len(a)):
        current = a[i]
        if not np.isnan(current) and tmp > 0:
            GapList.append(tmp)
            tmp = 0
        if np.isnan(current):
            tmp = tmp + 1
    return GapList

# Create a numpy array with some NaN values to test the function
a = np.array([1.0, np.nan, np.nan, 2.0, 3.0, np.nan, 4.0, np.nan, np.nan, np.nan, 5.0])

ARGS = {
    'a': a
}
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 list, ndim 1, dtype family int — 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
createAngleDatabase Ran
ran on a synthesized input
alexhang212/Dataset-3DPOP/POP3D_AP/ApplicationExamples/ComputePlanes.py
code served (permissive licence) · get_code("1187036ef80b21e1")
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 torch
import pandas as pd

torch.manual_seed(0)

# The function expects customObjects to be iterable
# It converts each feature to a string and appends '_x', '_y', '_z'
# So we need a list/tuple of objects that can be converted to strings

customObjects = ['joint_0', 'joint_1', 'joint_2']

ARGS = {
    'customObjects': customObjects
}
Setup block drafted by
a language model, Claude Haiku 4.5, prompt python_fixture_v1, through the batch API
Imports
pandas, 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: none by the harness; any seed is in the setup block
getColor Ran
ran on a synthesized input
alexhang212/Dataset-3DPOP/POP3D_Reader/ImageReader.py
code served (permissive licence) · get_code("47e806fbb4e95458")
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, ndim 1, dtype family int; 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

# The function takes a single parameter: keyPoint (a string)
# It checks if the keyPoint string ends with various suffixes and returns colors accordingly

keyPoint_examples = [
    'bird_beak',
    'bird_nose',
    'leftEye',
    'rightEye',
    'leftShoulder',
    'rightShoulder',
    'topKeel',
    'bottomKeel',
    'tail',
    'unknown'
]

# Create ARGS dict with a valid keyPoint string
ARGS = {
    'keyPoint': 'bird_beak'
}
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 tuple, ndim 1, dtype family int — a tuple 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
getFeaturesInCameraSpace Ran
ran on a synthesized input
alexhang212/Dataset-3DPOP/POP3D_AP/ApplicationExamples/automaticAnnotationTool.py
code served (permissive licence) · get_code("345c6cfc43587c78")
How it ran

Ran in Syntology's sandbox on this synthesized input (drafted by a language model, Claude Haiku 4.5, prompt python_fixture_v1, through the batch API) and returned bool, ndim 0, dtype family float, shape () as a later run recorded it. The image it ran in was not recorded. If your copy returns the same type, ndim and shape on these inputs, it matched what we observed; nothing here checks that. It does NOT mean the output is correct: nothing has compared it to the paper.

import torch
import numpy as np

torch.manual_seed(0)

class MockViconCamObject:
    def __init__(self):
        self.features = None
    
    def clearFeatures(self):
        self.features = None
    
    def transferFeaturesToObjectSpace(self, featureDictViconSpace):
        # Transform features from Vicon space to camera/object space
        # Return a modified copy of the feature dict
        featureDictCamSpace = {}
        for key, value in featureDictViconSpace.items():
            if isinstance(value, torch.Tensor):
                # Simple mock transformation: add a constant offset
                featureDictCamSpace[key] = value + 1.0
            else:
                featureDictCamSpace[key] = value
        return featureDictCamSpace
    
    def setFeatures(self, featureDictCamSpace):
        self.features = featureDictCamSpace

# Create mock Vicon camera objects
viconCamObjects = [
    MockViconCamObject(),
    MockViconCamObject(),
]

# Create feature dict in Vicon space with small tensors
featureDictViconSpace = {
    'position': torch.randn(2, 4),
    'velocity': torch.randn(2, 3),
    'metadata': 'test',
}

ARGS = {
    'viconCamObjects': viconCamObjects,
    'featureDictViconSpace': featureDictViconSpace,
}
Setup block drafted by
a language model, Claude Haiku 4.5, prompt python_fixture_v1, through the batch API
Imports
numpy, torch — import names as written, not package names to install; a name the repository itself provides is not a package
Returned
type bool, ndim 0, dtype family float, shape () — shape () as a later run recorded it
dtype family
int means an integer dtype; float means any other dtype the instrument saw, bool and complex included (fixture_check.py records only these two families)
Environment
image not recorded; re-verify against syntology-harvest-sandbox:v5-deps; not recorded for this sample: the harness installs a cuda-to-cpu shim (cuda_shim.py) before every sample and no run records whether a sample needed it; seed: none by the harness; any seed is in the setup block
getFeaturesInViconSpace Ran
ran on a synthesized input
alexhang212/Dataset-3DPOP/POP3D_AP/ApplicationExamples/automaticAnnotationTool.py
code served (permissive licence) · get_code("414809e810f2efeb")
How it ran

Ran in Syntology's sandbox on this synthesized input (drafted by a language model, Claude Haiku 4.5, prompt python_fixture_v1, through the batch API) and returned dict. The image it ran in was not recorded. If your copy returns the same type on these inputs, it matched what we observed; nothing here checks that. It does NOT mean the output is correct: nothing has compared it to the paper.

import torch
import numpy as np

class ViconObject:
    def __init__(self, name):
        self.name = name
        self.transformation_rotation = None
        self.transformation_translation = None
        self.features = {f"{name}_feature_1": torch.randn(2, 4), f"{name}_feature_2": torch.randn(2, 3)}
    
    def __getattribute__(self, attr):
        return object.__getattribute__(self, attr)
    
    def setTransformationParameters(self, rotation, translation):
        self.transformation_rotation = rotation
        self.transformation_translation = translation
    
    def transferFeaturesToViconSpace(self):
        return {k: v for k, v in self.features.items()}

torch.manual_seed(0)

vicon_obj_1 = ViconObject("object_1")
vicon_obj_2 = ViconObject("object_2")

viconObjects = [vicon_obj_1, vicon_obj_2]

transformationParamDict = {
    "object_1_validity": True,
    "object_1_rotation": np.array([[1.0, 0.0, 0.0], [0.0, 1.0, 0.0], [0.0, 0.0, 1.0]]),
    "object_1_translation": np.array([0.1, 0.2, 0.3]),
    "object_2_validity": False,
    "object_2_rotation": np.array([[1.0, 0.0, 0.0], [0.0, 1.0, 0.0], [0.0, 0.0, 1.0]]),
    "object_2_translation": np.array([0.0, 0.0, 0.0]),
}

ARGS = {
    "viconObjects": viconObjects,
    "transformationParamDict": transformationParamDict,
}
Setup block drafted by
a language model, Claude Haiku 4.5, prompt python_fixture_v1, through the batch API
Imports
numpy, torch — import names as written, not package names to install; a name the repository itself provides is not a package
Returned
type 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
getFrameData Ran
ran on a synthesized input
alexhang212/Dataset-3DPOP/POP3D_AP/ApplicationExamples/ReprojectCustomFeatures.py
code served (permissive licence) · get_code("5a3bb34299738248")
How it ran

Ran in Syntology's sandbox on this synthesized input (drafted by a language model, Claude Haiku 4.5, prompt python_fixture_v1, through the batch API) and returned dict. The image it ran in was not recorded. If your copy returns the same type on these inputs, it matched what we observed; nothing here checks that. It does NOT mean the output is correct: nothing has compared it to the paper.

import pandas as pd
import numpy as np

# Create a mock dataframe object with a 'data' attribute
class MockDataFrame:
    def __init__(self):
        # Create a pandas DataFrame with frame column and feature data
        # Structure: frame | feature1_x | feature1_y | feature1_z | feature2_x | feature2_y | feature2_z
        self.data = pd.DataFrame({
            'frame': [0, 4, 8],
            'feature1_x': [1.0, 2.0, 3.0],
            'feature1_y': [1.1, 2.1, 3.1],
            'feature1_z': [1.2, 2.2, 3.2],
            'feature2_x': [4.0, 5.0, 6.0],
            'feature2_y': [4.1, 5.1, 6.1],
            'feature2_z': [4.2, 5.2, 6.2],
        })

df = MockDataFrame()

# dataFrameNumber: query frame number (will be multiplied by 2 internally)
# If we pass 2, it will query for frame == 4, which exists in our test data
dataFrameNumber = 2

# FeatureList: list of feature names
# Must have length such that len(FeatureList) * 3 == len(subDataList)
# subDataList has 6 elements (2 features × 3 coordinates each)
# So FeatureList should have 2 elements
FeatureList = ['feature1', 'feature2']

ARGS = {
    'df': df,
    'dataFrameNumber': dataFrameNumber,
    'FeatureList': FeatureList,
}
Setup block drafted by
a language model, Claude Haiku 4.5, prompt python_fixture_v1, through the batch API
Imports
numpy, pandas — 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
getRoiPoint Ran
ran on a synthesized input
alexhang212/Dataset-3DPOP/POP3D_AP/ApplicationExamples/ManualAnnot.py
code served (permissive licence) · get_code("4217b05b7b44bd93")
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 list, ndim 1, dtype family int; 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

torch.manual_seed(0)

# Construct arguments for getRoiPoint
point = [150.5, 200.7]  # A point with x=150, y=200
height = 600
width = 800
size = 200

ARGS = {
    "point": point,
    "height": height,
    "width": width,
    "size": size
}
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 list, ndim 1, dtype family int — 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
videoWriter Ran
ran on a synthesized input
alexhang212/Dataset-3DPOP/POP3D_AP/ApplicationExamples/automaticAnnotationTool.py
code served (permissive licence) · get_code("d86f67d9643cb8b4")
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 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 cv2 as cv

# The function iterates over windowNames with len() and indexing,
# concatenates strings, and passes dimensions to cv.VideoWriter
# We need a list of window name strings

windowNames = ['camera1_', 'camera2_']
FPS = 30
imageWidth = 1920
imageHeight = 1080

ARGS = {
    'windowNames': windowNames,
    'FPS': FPS,
    'imageWidth': imageWidth,
    'imageHeight': imageHeight
}
Setup block drafted by
a language model, Claude Haiku 4.5, prompt python_fixture_v1, through the batch API
Imports
cv2 — 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
PlotLine Not yet run
unverified (harvested, no recorded run)
alexhang212/Dataset-3DPOP/POP3D_AP/ApplicationExamples/ManualAnnot.py
code served (permissive licence) · get_code("2a96de96e55f2fa0")
How it ran

no invocation recorded.

getType Not yet run
unverified (harvested, no recorded run)
alexhang212/Dataset-3DPOP/Util/VisualizeUtil.py
code served (permissive licence) · get_code("de4617b55f0ad1b1")
How it ran

no invocation recorded.

Also present in other papers' code

Of 15 samples lifted from this paper's linked repositories, 15 carry a function identity and 0 do not. Of the 15 distinct functions with an identity, 0 are also present in other papers whose code we hold (0 with at least one paper dated earlier, 0 the same month, 0 later) and 15 are only here among the 52,977 papers whose code Syntology holds (as of 2026-09-26).

Only here means: among the 52,977 papers whose code Syntology holds (51,887 with lifted samples, 1,090 through their assemblies only), as of 2026-09-26, no other paper's lifted samples or current assembly bricks contain an identical function body after normalisation.

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

Recent advances in machine learning and computer vision are revolutionizing the field of animal behavior by enabling researchers to track the poses and locations of freely moving animals without any marker attachment. However, large datasets of annotated images of animals for markerless pose tracking, especially high-resolution images taken from multiple angles with accurate 3D annotations, are still scant. Here, we propose a method that uses a motion capture (mo-cap) system to obtain a large amount of annotated data on animal movement and posture (2D and 3D) in a semi-automatic manner. Our method is novel in that it extracts the 3D positions of morphological keypoints (e.g eyes, beak, tail) in reference to the positions of markers attached to the animals. Using this method, we obtained, and offer here, a new dataset -3D-POP with approximately 300k annotated frames (4 million instances) in the form of videos having groups of one to ten freely moving birds from 4 different camera views in a 3.6m x 4.2m area. 3D-POP is the first dataset of flocking birds with accurate keypoint annotations in 2D and 3D along with bounding box and individual identities and will facilitate the development of solutions for problems of 2D to 3D markerless pose, trajectory tracking, and identification in birds.

For agents

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

get_harvested_code_for_paper("2303.13174")
get_code_for_paper("2303.13174")
have("2303.13174")

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

curl https://syntology.ai/api/ran/2303.13174.json

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

[![Syntology run record](https://syntology.ai/api/ran/2303.13174.svg)](https://syntology.ai/paper/2303.13174)

Connect an agent — have() is free.