SYNTOLOGY HomeExplorerAtlasCodeMethodologyAboutDevelopersFeedPricing
Paper · aaai_19899 · AAAI · 2022

Keypoint Message Passing for Video-Based Person Re-Identification

Bernt Schiele, Jian Yang, Juergen Gall, Shanshan Zhang, Di Chen, Andreas Döring

Open in the Atlas

Code that ran

7 of 7 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 6 have no contract verdict; 1 failure of our instrument; 0 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/aaai_19899.json · badge: /api/ran/aaai_19899.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
DeanChan/KeypointMessagePassing canonical (classifier label)
classifier basis: census
7 ran · 0 no recorded run
FunctionStatusWhere it lives
imagedata_kwargs Ran
ran on a synthesized input
DeanChan/KeypointMessagePassing/configs/default_config.py
code served (permissive licence) · get_code("8cd256e691f38e66")
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 torch.nn as nn

torch.manual_seed(0)

class Config:
    def __init__(self):
        self.data = DataConfig()
        self.train = TrainConfig()
        self.test = TestConfig()
        self.sampler = SamplerConfig()
        self.cuhk03 = CUHK03Config()
        self.market1501 = Market1501Config()
        self.use_gpu = False

class DataConfig:
    def __init__(self):
        self.root = '/data'
        self.sources = ['market1501']
        self.targets = ['dukemtmc']
        self.height = 256
        self.width = 128
        self.transforms = None
        self.norm_mean = [0.485, 0.456, 0.406]
        self.norm_std = [0.229, 0.224, 0.225]
        self.split_id = 0
        self.combineall = False
        self.load_train_targets = False
        self.workers = 4
        self.batch_size_train = 32

class TrainConfig:
    def __init__(self):
        self.batch_size = 32

class TestConfig:
    def __init__(self):
        self.batch_size = 128

class SamplerConfig:
    def __init__(self):
        self.num_instances = 4
        self.train_sampler = 'RandomSampler'
        self.train_sampler_t = 'RandomSampler'

class CUHK03Config:
    def __init__(self):
        self.labeled_images = True
        self.classic_split = True

class Market1501Config:
    def __init__(self):
        self.use_500k_distractors = False

cfg = Config()

ARGS = {
    'cfg': cfg
}
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 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
map_joint_dict Ran
ran on a synthesized input
DeanChan/KeypointMessagePassing/lib/utils/pose_visualization.py
code served (permissive licence) · get_code("512cac7d1c002911")
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

torch.manual_seed(0)

# The function expects 'joints' to be a 2D array-like object with shape (N, 2+)
# where it accesses joints[i][0] and joints[i][1] for x and y coordinates
# It iterates through all rows and converts to int

joints = torch.tensor([
    [10.5, 20.3],
    [30.7, 40.1],
], dtype=torch.float32)

ARGS = {
    "joints": joints
}
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
null_collate_fn Ran
ran on a synthesized input
DeanChan/KeypointMessagePassing/lib/utils/tools.py
code served (permissive licence) · get_code("f6ef2cf0ed9b8e69")
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. 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

torch.manual_seed(0)

# null_collate_fn simply returns its input unchanged
# The function takes a single parameter 'x' and returns it as-is
# x should be some kind of batch data - typically a list of samples

# Create a simple batch of data - a list of tuples (features, labels)
batch_data = [
    (torch.randn(4), torch.tensor(0)),
    (torch.randn(4), torch.tensor(1)),
]

ARGS = {
    "x": batch_data
}
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 list — a list 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
posedata_kwargs Ran
ran on a synthesized input
DeanChan/KeypointMessagePassing/configs/default_config.py
code served (permissive licence) · get_code("7e40a65f0d9d1790")
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 torch.nn as nn

torch.manual_seed(0)

class Config:
    def __init__(self):
        self.data = DataConfig()
        self.train = TrainConfig()
        self.test = TestConfig()
        self.sampler = SamplerConfig()
        self.video = VideoConfig()
        self.use_gpu = False

class DataConfig:
    def __init__(self):
        self.root = '/data/pose'
        self.sources = ['source1', 'source2']
        self.targets = ['target1']
        self.height = 256
        self.width = 256
        self.transforms = None
        self.split_id = 0
        self.combineall = False
        self.workers = 4
        self.headless = False
        self.rm_glitches = True
        self.include_dynamics = True
        self.include_spatial_links = True
        self.include_temporal_links = True
        self.pose_w_img = True

class TrainConfig:
    def __init__(self):
        self.batch_size = 32

class TestConfig:
    def __init__(self):
        self.batch_size = 16

class SamplerConfig:
    def __init__(self):
        self.num_instances = 4
        self.train_sampler = 'random'

class VideoConfig:
    def __init__(self):
        self.seq_len = 8
        self.sample_method = 'uniform'

cfg = Config()

ARGS = {
    'cfg': cfg
}
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 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
read_json Ran
ran; our contract draft was wrong, not the code
DeanChan/KeypointMessagePassing/lib/utils/serialization.py
pointer only (licence: MIT); the stored copy get_code serves, from luomingshuang/M3GPT, is pointer only (licence: NONE) · get_code("7d6e263caea8e963")
How it ran

The body of this sample is withheld: this paper's link allows the body inline (MIT), but get_code serves the stored copy, first harvested from luomingshuang/M3GPT, 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.

read_keypoints Ran
ran on a synthesized input
DeanChan/KeypointMessagePassing/lib/utils/tools.py
code served (permissive licence) · get_code("b43c3385f983af1f")
How it ran

Ran in Syntology's sandbox on this synthesized input (drafted by a language model, Claude Haiku 4.5, prompt python_fixture_v1, through the batch API) and returned ndarray, ndim 2, dtype family float, shape (3, 4) as a later run recorded it. The image it ran in was not recorded. If your copy returns the same type, ndim and shape on these inputs, it matched what we observed; nothing here checks that. It does NOT mean the output is correct: nothing has compared it to the paper.

import os
import os.path as osp
import numpy as np
import tempfile

# Create a temporary file with keypoints data
temp_dir = tempfile.mkdtemp()
keypoints_file = osp.join(temp_dir, "keypoints.txt")

# Write some sample keypoint data to the file
# The function expects the file to contain a line that can be eval'd into a numpy array
sample_keypoints = [[1.5, 2.3, 3.1, 4.2], [5.1, 6.2, 7.3, 8.4], [9.5, 10.1, 11.2, 12.3]]
with open(keypoints_file, 'wb') as f:
    f.write(str(sample_keypoints).encode())

ARGS = {
    'path': keypoints_file
}
Setup block drafted by
a language model, Claude Haiku 4.5, prompt python_fixture_v1, through the batch API
Imports
numpy — import names as written, not package names to install; a name the repository itself provides is not a package
Returned
type ndarray, ndim 2, dtype family float, shape (3, 4) — shape (3, 4) as a later run recorded it
dtype family
int means an integer dtype; float means any other dtype the instrument saw, bool and complex included (fixture_check.py records only these two families)
Environment
image not recorded; re-verify against syntology-harvest-sandbox:v5-deps; not recorded for this sample: the harness installs a cuda-to-cpu shim (cuda_shim.py) before every sample and no run records whether a sample needed it; seed: none by the harness; any seed is in the setup block
videodata_kwargs Ran
ran on a synthesized input
DeanChan/KeypointMessagePassing/configs/default_config.py
code served (permissive licence) · get_code("0951e0d511f67f90")
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 torch.nn as nn

torch.manual_seed(0)

class MockTransform:
    def __call__(self, x):
        return x

class MockConfig:
    def __init__(self):
        self.data = MockDataConfig()
        self.train = MockTrainConfig()
        self.test = MockTestConfig()
        self.sampler = MockSamplerConfig()
        self.video = MockVideoConfig()
        self.use_gpu = False

class MockDataConfig:
    def __init__(self):
        self.root = "/tmp/data"
        self.sources = ["camera1", "camera2"]
        self.targets = ["action", "person"]
        self.height = 224
        self.width = 224
        self.transforms = MockTransform()
        self.norm_mean = [0.485, 0.456, 0.406]
        self.norm_std = [0.229, 0.224, 0.225]
        self.split_id = 0
        self.combineall = False
        self.workers = 4

class MockTrainConfig:
    def __init__(self):
        self.batch_size = 32

class MockTestConfig:
    def __init__(self):
        self.batch_size = 16

class MockSamplerConfig:
    def __init__(self):
        self.num_instances = 4
        self.train_sampler = "random"

class MockVideoConfig:
    def __init__(self):
        self.seq_len = 8
        self.sample_method = "uniform"

cfg = MockConfig()

ARGS = {
    "cfg": cfg
}
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 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 7 samples lifted from this paper's linked repositories, 7 carry a function identity and 0 do not. Of the 7 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, 1 later) and 5 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
read_json 8 other papers
4 dated earlier, 3 later, 1 same year, order not determinable
1612.05601 (2016-12)
SonoNet: Real-Time Detection and Localisation of Fetal Standard Scan Planes in Freehand Ultrasound
1903.11306 (2019-03)
Linkage Based Face Clustering via Graph Convolution Network
2001.04193 (2020-01)
Deep Learning for Person Re-identification: A Survey and Outlook
and 5 more not named here
videodata_kwargs 4 other papers
3 dated earlier, 1 same year, order not determinable
1910.10093 (2019-10)
Torchreid: A Library for Deep Learning Person Re-Identification in Pytorch
2010.05435 (2020-10)
Top-DB-Net: Top DropBlock for Activation Enhancement in Person Re-Identification
Zhao_PhD_Learning_Learning_With_Pompeiu-Hausdorff_Distances_for_Video-Based_Vehicle_Re-Identification_CVPR_2021_paper (2021 (year only))
PhD Learning: Learning with Pompeiu-hausdorff Distances for Video-based Vehicle Re-Identification
and 1 more not named here

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

Video-based person re-identification (re-ID) is an important technique in visual surveillance systems which aims to match video snippets of people captured by different cameras. Existing methods are mostly based on convolutional neural networks (CNNs), whose building blocks either process local neighbor pixels at a time, or, when 3D convolutions are used to model temporal information, suffer from the misalignment problem caused by person movement. In this paper, we propose to overcome the limitations of normal convolutions with a human-oriented graph method. Specifically, features located at person joint keypoints are extracted and connected as a spatial-temporal graph. These keypoint features are then updated by message passing from their connected nodes with a graph convolutional network (GCN). During training, the GCN can be attached to any CNN-based person re-ID model to assist representation learning on feature maps, whilst it can be dropped after training for better inference speed. Our method brings significant improvements over the CNN-based baseline model on the MARS dataset with generated person keypoints and a newly annotated dataset: PoseTrackReID. It also defines a new state-of-the-art method in terms of top-1 accuracy and mean average precision in comparison to prior works.

For agents

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

get_harvested_code_for_paper("aaai_19899")
get_code_for_paper("aaai_19899")
have("aaai_19899")

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

curl https://syntology.ai/api/ran/aaai_19899.json

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

[![Syntology run record](https://syntology.ai/api/ran/aaai_19899.svg)](https://syntology.ai/paper/aaai_19899)

Connect an agent — have() is free.