SYNTOLOGY HomeExplorerAtlasCodeMethodologyAboutDevelopersFeedPricing
Paper · 2007.09389 · 2020

SRNet: Improving Generalization in 3D Human Pose Estimation with a Split-and-Recombine Approach

arXiv · PDF · Open in the Atlas

Code that ran

6 of 9 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; 1 failure of our instrument; 3 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/2007.09389.json · badge: /api/ran/2007.09389.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
ailingzengzzz/Split-and-Recombine-Net canonical (classifier label)
classifier basis: census
6 ran · 3 no recorded run
FunctionStatusWhere it lives
image_coordinates Ran
ran on a synthesized input
ailingzengzzz/Split-and-Recombine-Net/common/transformation/cam_utils.py
code served (permissive licence) · get_code("2c8338c89fe8975c")
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 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 torch

torch.manual_seed(0)

# Create a 3D tensor with shape (2, 4, 3) - batch of 2, 4 spatial points, 3 channels (x, y, z)
# Values in normalized range roughly [-1, 1] for x, y and [0, 1] for z
X_3d = torch.randn(2, 4, 3)

# Width and height parameters
w = 512
h = 384

ARGS = {
    "X": X_3d,
    "w": w,
    "h": h
}
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 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
mpjae Ran
ran on a synthesized input
ailingzengzzz/Split-and-Recombine-Net/common/common_pytorch/loss/loss_family.py
code served (permissive licence) · get_code("74396e4069c4e8d6")
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 0, dtype family float, shape () as a later run recorded it. 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, 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 computes mean per-joint angle error between predicted and target
# Both must be tensors with the same shape
# The function expects 3D bone vector angle errors
# Typical shape would be (batch_size, num_joints) or similar

# Create small tensors: batch_size=2, num_joints=4
predicted = torch.randn(2, 4)
target = torch.randn(2, 4)

ARGS = {
    "predicted": predicted,
    "target": target,
}
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 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
mpjpe Ran
ran; our contract draft was wrong, not the code
ailingzengzzz/Split-and-Recombine-Net/common/common_pytorch/loss/loss_family.py
pointer only (licence: Apache-2.0); the stored copy get_code serves, from davidpengucf/DAF-DG, is pointer only (licence: NONE) · get_code("3afb541a8147d123")
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 davidpengucf/DAF-DG, 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.

mpjpe_smooth Ran
ran on a synthesized input
ailingzengzzz/Split-and-Recombine-Net/common/common_pytorch/loss/loss_family.py
code served (permissive licence) · get_code("b72379b0ad4e784a")
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 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)

# Create small tensors for predicted and target
# Shape: (batch_size=2, spatial_dims=4)
predicted = torch.randn(2, 4)
target = torch.randn(2, 4)

# threshold: a scalar value for the smoothing threshold
threshold = 0.5

# mi: a scalar exponent parameter for the smooth loss
mi = 0.5

# L1: boolean flag for whether to use L1 distance or squared difference
L1 = False

ARGS = {
    "predicted": predicted,
    "target": target,
    "threshold": threshold,
    "mi": mi,
    "L1": L1,
}
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 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
normalize_screen_coordinates Ran
ran on a synthesized input
ailingzengzzz/Split-and-Recombine-Net/common/transformation/cam_utils.py
code served (permissive licence) · get_code("e53a897a491c4627")
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

# The function normalizes screen coordinates given width and height
# X can be 2D or 3D coordinates (last dimension is 2 or 3)
# w is width (scalar), h is height (scalar)

# Create sample coordinate data with shape (2, 4, 3) - batch of 2, 4 points, 3 coords (x, y, z)
X = np.random.rand(2, 4, 3).astype(np.float32) * 100

# Width and height of screen
w = 640
h = 480

ARGS = {
    "X": X,
    "w": w,
    "h": h
}
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
process_3d Ran
ran on a synthesized input
ailingzengzzz/Split-and-Recombine-Net/common/transformation/aug_rotate.py
code served (permissive licence) · get_code("44aeda9bcbffef2a")
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 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 torch

torch.manual_seed(0)

# The function process_3d takes pose_3d_in and subtracts the first joint position
# from all joints to center the pose. It expects a tensor with shape
# (batch_size, num_joints, num_dimensions)

pose_3d_in = torch.randn(2, 4, 3)  # batch=2, num_joints=4, dimensions=3

ARGS = {
    "pose_3d_in": pose_3d_in,
}
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 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
axis_rotation Not yet run
unverified (harvested, no recorded run)
ailingzengzzz/Split-and-Recombine-Net/common/transformation/aug_rotate.py
code served (permissive licence) · get_code("be5ee7aae00d0f33")
How it ran

no invocation recorded.

rotate Not yet run
unverified (harvested, no recorded run)
ailingzengzzz/Split-and-Recombine-Net/common/transformation/aug_rotate.py
code served (permissive licence) · get_code("fb05d05fdb09695a")
How it ran

no invocation recorded.

wrap Not yet run
unverified (harvested, no recorded run)
ailingzengzzz/Split-and-Recombine-Net/common/common_pytorch/utils.py
code served (permissive licence) · get_code("5323ac47f72a7095")
How it ran

no invocation recorded.

Also present in other papers' code

Of 9 samples lifted from this paper's linked repositories, 9 carry a function identity and 0 do not. Of the 9 distinct functions with an identity, 4 are also present in other papers whose code we hold (2 with at least one paper dated earlier, 0 the same month, 4 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
mpjpe 4 other papers
1 dated earlier, 3 later
1904.03345 (2019-04)
Semantic Graph Convolutional Networks for 3D Human Pose Regression
2211.16940 (2022-11)
DiffPose: Toward More Reliable 3D Pose Estimation
2308.10436 (2023-08)
Approximately Equivariant Graph Networks
and 1 more not named here
wrap 2 other papers
1 dated earlier, 1 later
2004.13985 (2020-04)
Motion Guided 3D Pose Estimation from Videos
2310.16288 (2023-10)
MotionAGFormer: Enhancing 3D Human Pose Estimation with a Transformer-GCNFormer Network
image_coordinates 1 other paper
1 later
2203.08713 (2022-03)
DeciWatch: A Simple Baseline for 10x Efficient 2D and 3D Pose Estimation
normalize_screen_coordinates 1 other paper
1 later
2203.08713 (2022-03)
DeciWatch: A Simple Baseline for 10x Efficient 2D and 3D Pose 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 2026-11-21.

Repositories linked to this paper

Some links come from the archived Papers with Code dataset (CC BY-SA 4.0): attribution and licence.

Abstract

Human poses that are rare or unseen in a training set are challenging for a network to predict. Similar to the long-tailed distribution problem in visual recognition, the small number of examples for such poses limits the ability of networks to model them. Interestingly, local pose distributions suffer less from the long-tail problem, i.e., local joint configurations within a rare pose may appear within other poses in the training set, making them less rare. We propose to take advantage of this fact for better generalization to rare and unseen poses. To be specific, our method splits the body into local regions and processes them in separate network branches, utilizing the property that a joint position depends mainly on the joints within its local body region. Global coherence is maintained by recombining the global context from the rest of the body into each branch as a low-dimensional vector. With the reduced dimensionality of less relevant body areas, the training set distribution within network branches more closely reflects the statistics of local poses instead of global body poses, without sacrificing information important for joint inference. The proposed split-and-recombine approach, called SRNet, can be easily adapted to both single-image and temporal models, and it leads to appreciable improvements in the prediction of rare and unseen poses.

For agents

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

get_harvested_code_for_paper("2007.09389")
get_code_for_paper("2007.09389")
have("2007.09389")

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

curl https://syntology.ai/api/ran/2007.09389.json

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

[![Syntology run record](https://syntology.ai/api/ran/2007.09389.svg)](https://syntology.ai/paper/2007.09389)

Connect an agent — have() is free.