SYNTOLOGY HomeExplorerAtlasCodeMethodologyAboutDevelopersFeedPricing
Paper · 2212.14474 · 2022

Learning 3D Human Pose Estimation from Dozens of Datasets using a Geometry-Aware Autoencoder to Bridge Between Skeleton Formats

arXiv · PDF · Open in the Atlas

Code that ran

5 of 8 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, 1 violated it and 3 have no contract verdict; 1 failure of our instrument; 3 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/2212.14474.json · badge: /api/ran/2212.14474.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
isarandi/metrabs canonical (classifier label)
classifier basis not recorded on the link
3 ran · 3 no recorded run
isarandi/posepile community (classifier label)
classifier basis not recorded on the link
2 ran · 0 no recorded run
FunctionStatusWhere it lives
fixed_padding_layer Ran
ran; our contract draft was wrong, not the code
isarandi/metrabs/metrabs_pytorch/backbones/efficientnet.py
code served (permissive licence) · get_code("2f6a6eccfb577d4a")
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 ZeroPad2d. 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

torch.manual_seed(0)

ARGS = {
    "kernel_size": 3,
    "rate": 1,
    "shifts": (0, 0)
}
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 ZeroPad2d — a ZeroPad2d output has no array shape of its own; none is stated — the instrument coerces any output to an array; for a ZeroPad2d 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
get_joint2bone_mat Ran
ran on a synthesized input
isarandi/posepile/src/posepile/joint_info.py
code served (permissive licence) · get_code("65506ac65d5fefd7")
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, 5) as a later run recorded it. The image it ran in was not recorded. If your copy returns the same type, ndim and shape on these inputs, it matched what we observed; nothing here checks that. It does NOT mean the output is correct: nothing has compared it to the paper.

import numpy as np

class JointInfo:
    def __init__(self, n_joints, stick_figure_edges):
        self.n_joints = n_joints
        self.stick_figure_edges = stick_figure_edges

# Create a minimal joint_info object
# n_joints: number of joints
# stick_figure_edges: list of (joint_index1, joint_index2) tuples defining bones
joint_info = JointInfo(
    n_joints=5,
    stick_figure_edges=[
        (0, 1),
        (1, 2),
        (2, 3),
        (3, 4)
    ]
)

ARGS = {
    "joint_info": joint_info
}
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, 5) — shape (4, 5) as a later run recorded it
dtype family
int means an integer dtype; float means any other dtype the instrument saw, bool and complex included (fixture_check.py records only these two families)
Environment
image not recorded; re-verify against syntology-harvest-sandbox:v5-deps; not recorded for this sample: the harness installs a cuda-to-cpu shim (cuda_shim.py) before every sample and no run records whether a sample needed it; seed: none by the harness; any seed is in the setup block
get_z Ran
ran on a synthesized input
isarandi/posepile/src/posepile/compositing.py
code served (permissive licence) · get_code("9f844d1da50ec728")
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 float32, 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 numpy as np
import torch
import torch.nn as nn

class MockCamera:
    def world_to_camera(self, world_coords):
        # Transform world coordinates to camera coordinates
        # Return shape matching input with z-depth in last dimension
        batch_size = world_coords.shape[0]
        camera_coords = world_coords.copy() if isinstance(world_coords, np.ndarray) else world_coords.clone()
        # Shift z values to simulate camera transformation
        if isinstance(camera_coords, np.ndarray):
            camera_coords[..., 2] = camera_coords[..., 2] + 5.0
        else:
            camera_coords[..., 2] = camera_coords[..., 2] + 5.0
        return camera_coords

class MockExample:
    def __init__(self):
        self.camera = MockCamera()
        # world_coords shape: (batch_size, num_points, 3)
        self.world_coords = np.array([
            [[1.0, 2.0, 3.0], [4.0, 5.0, 6.0]],
            [[7.0, 8.0, 9.0], [10.0, 11.0, 12.0]]
        ], dtype=np.float32)

ex = MockExample()

ARGS = {
    "ex": ex
}
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 float32, 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
rearrange_tf_to_pt Ran
ran, violated the contract we drafted
isarandi/metrabs/metrabs_pytorch/convert_model_from_tf.py
code served (permissive licence) · get_code("697f5aa8d9395594")
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, shape (5, 4, 2, 3) as a later run recorded it. The image it ran in was not recorded. If your copy returns the same type, ndim and shape on these inputs, it matched what we observed; nothing here checks that. It does NOT mean the output is correct: nothing has compared it to the paper.

import torch
import einops

torch.manual_seed(0)

# Test case 1: 4D tensor with depthwise=False
value_4d = torch.randn(2, 3, 4, 5)  # h=2, w=3, c_in=4, c_out=5

# Test case 2: 4D tensor with depthwise=True
value_4d_dw = torch.randn(2, 3, 4, 5)  # h=2, w=3, c_in=4, c_out=5

# Test case 3: 2D tensor
value_2d = torch.randn(4, 5)  # c_in=4, c_out=5

# Test case 4: 1D tensor
value_1d = torch.randn(4)  # c_in=4

ARGS = {
    "value": value_4d,
    "depthwise": False
}
Setup block drafted by
a language model, Claude Haiku 4.5, prompt python_fixture_v1, through the batch API
Imports
einops, 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, shape (5, 4, 2, 3) — shape (5, 4, 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
to_homogeneous Ran
ran on a synthesized input
isarandi/metrabs/metrabs_pytorch/ptu3d.py
code served (permissive licence) · get_code("7766e0fe4266257a")
How it ran

Ran in Syntology's sandbox on this synthesized input (drafted by a language model, Claude Haiku 4.5, prompt python_fixture_v1, through the batch API) and returned Tensor, ndim 2, dtype family float, shape (2, 5) as a later run recorded it. The image it ran in was not recorded. If your copy returns the same type, ndim and shape on these inputs, it matched what we observed; nothing here checks that. It does NOT mean the output is correct: nothing has compared it to the paper.

import torch

torch.manual_seed(0)

# The function to_homogeneous takes a tensor x and concatenates ones along the last dimension
# It uses:
# - x[..., :1] to get the last dimension (slice of size 1)
# - torch.ones_like() to create a tensor of ones with same shape as x[..., :1]
# - torch.cat() to concatenate along dim=-1
# 
# x should be a tensor with at least 1 dimension

# Create a small tensor with distinct dimensions: batch=2, features=4
x = torch.randn(2, 4)

ARGS = {
    "x": x
}
Setup block drafted by
a language model, Claude Haiku 4.5, prompt python_fixture_v1, through the batch API
Imports
torch — import names as written, not package names to install; a name the repository itself provides is not a package
Returned
type Tensor, ndim 2, dtype family float, shape (2, 5) — shape (2, 5) as a later run recorded it
dtype family
int means an integer dtype; float means any other dtype the instrument saw, bool and complex included (fixture_check.py records only these two families)
Environment
image not recorded; re-verify against syntology-harvest-sandbox:v5-deps; not recorded for this sample: the harness installs a cuda-to-cpu shim (cuda_shim.py) before every sample and no run records whether a sample needed it; seed: none by the harness; any seed is in the setup block
mean_stdev_masked Not yet run
unverified (harvested, no recorded run)
isarandi/metrabs/metrabs_pytorch/ptu.py
code served (permissive licence) · get_code("a019fdb41511cf6c")
How it ran

no invocation recorded.

reduce_mean_masked Not yet run
unverified (harvested, no recorded run)
isarandi/metrabs/metrabs_pytorch/ptu.py
code served (permissive licence) · get_code("11ac9049df0417ee")
How it ran

no invocation recorded.

reduce_sum_masked Not yet run
unverified (harvested, no recorded run)
isarandi/metrabs/metrabs_pytorch/ptu.py
code served (permissive licence) · get_code("c4919f8033efe728")
How it ran

no invocation recorded.

Also present in other papers' code

Of 8 samples lifted from this paper's linked repositories, 8 carry a function identity and 0 do not. Of the 8 distinct functions with an identity, 2 are also present in other papers whose code we hold (0 with at least one paper dated earlier, 0 the same month, 2 later) and 6 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
fixed_padding_layer 1 other paper
1 later
2407.07532 (2024-07)
Neural Localizer Fields for Continuous 3D Human Pose and Shape Estimation
rearrange_tf_to_pt 1 other paper
1 later
2407.07532 (2024-07)
Neural Localizer Fields for Continuous 3D Human Pose and Shape Estimation

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

Deep learning-based 3D human pose estimation performs best when trained on large amounts of labeled data, making combined learning from many datasets an important research direction. One obstacle to this endeavor are the different skeleton formats provided by different datasets, i.e., they do not label the same set of anatomical landmarks. There is little prior research on how to best supervise one model with such discrepant labels. We show that simply using separate output heads for different skeletons results in inconsistent depth estimates and insufficient information sharing across skeletons. As a remedy, we propose a novel affine-combining autoencoder (ACAE) method to perform dimensionality reduction on the number of landmarks. The discovered latent 3D points capture the redundancy among skeletons, enabling enhanced information sharing when used for consistency regularization. Our approach scales to an extreme multi-dataset regime, where we use 28 3D human pose datasets to supervise one model, which outperforms prior work on a range of benchmarks, including the challenging 3D Poses in the Wild (3DPW) dataset. Our code and models are available for research purposes.

For agents

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

get_harvested_code_for_paper("2212.14474")
get_code_for_paper("2212.14474")
have("2212.14474")

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

curl https://syntology.ai/api/ran/2212.14474.json

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

[![Syntology run record](https://syntology.ai/api/ran/2212.14474.svg)](https://syntology.ai/paper/2212.14474)

Connect an agent — have() is free.