SYNTOLOGY HomeExplorerAtlasCodeMethodologyAboutDevelopersFeedPricing
Paper · 2104.01797 · CVPR · 2021

Monocular 3D Multi-Person Pose Estimation by Integrating Top-Down and Bottom-Up Networks

Bo Yang, Yu Cheng, Bo Wang, Robby Tan

arXiv · PDF · Open in the Atlas

Code that ran

2 of 2 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; 1 honoured the contract we drafted, 0 violated it and 1 has no contract verdict; 0 failures 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/2104.01797.json · badge: /api/ran/2104.01797.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
3dpose/3D-Multi-Person-Pose community
Papers with Code archive
2 ran · 0 no recorded run
FunctionStatusWhere it lives
PoseMatcher Ran
ran on a synthesized input
3dpose/3D-Multi-Person-Pose/lib/posematcher.py
code served (permissive licence) · get_code("76fc9c1f86dda18f")
How it ran

Ran in Syntology's sandbox on this synthesized input (drafted by a language model, class construct-then-call prompt, model not recorded) and returned NoneType. 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
import pickle
import os
import tempfile
from collections import defaultdict

# Create temporary directories for input and output
temp_dir = tempfile.mkdtemp()
top_down_path = os.path.join(temp_dir, "top_down")
btm_up_path = os.path.join(temp_dir, "btm_up.pkl")
pts_out_path = os.path.join(temp_dir, "pts_out")
dep_out_path = os.path.join(temp_dir, "dep_out")
gt_dep_path = os.path.join(temp_dir, "gt_dep")

os.makedirs(top_down_path)
os.makedirs(gt_dep_path)

# Create mock top-down data (video results)
# Each video has frames with multiple people, each person has 15 joints with 3D coords
num_videos = 2
num_frames = 3
num_people_per_frame = 2
num_joints = 15

for vid_idx in range(1, num_videos + 1):
    frames_data = []
    for frame_idx in range(num_frames):
        people_in_frame = []
        for person_idx in range(num_people_per_frame):
            # Shape: (3, 15) - 3D coordinates for 15 joints
            person_pose = np.random.randn(3, num_joints).astype(np.float32)
            people_in_frame.append(person_pose)
        frames_data.append(people_in_frame)
    
    with open(os.path.join(top_down_path, f"{vid_idx}.pkl"), "wb") as f:
        pickle.dump(frames_data, f)

# Create mock bottom-up data
# Structure: dict[vid_idx][frame_idx] = (?, depths, ?, poses)
# depths: (num_detections, 1, 3) - depth values
# poses: (num_detections, 15, 3) - 3D poses transposed to (num_detections, 3, 15) when loaded
btm_up_data = {}
num_detections = 3

for vid_idx in range(1, num_videos + 1):
    btm_up_data[vid_idx] = []
    for frame_idx in range(num_frames):
        # Create mock data: (placeholder, depths, placeholder, poses)
        depths = np.random.randn(num_detections, 1, 3).astype(np.float32)
        # Poses in shape (num_detections, 15, 3) which will be transposed to (num_detections, 3, 15)
        poses = np.random.randn(num_detections, num_joints, 3).astype(np.float32)
        frame_data = (None, depths, None, poses)
        btm_up_data[vid_idx].append(frame_data)

with open(btm_up_path, "wb") as f:
    pickle.dump(btm_up_data, f)

# Create mock ground truth depth data
# Structure: one file per video with depth ground truth
for vid_idx in range(num_videos):
    # Shape: (num_people, num_frames) ground truth depths
    depth_gt = np.random.randn(num_people_per_frame, num_frames).astype(np.float32)
    with open(os.path.join(gt_dep_path, f"{vid_idx}_depth.pkl"), "wb") as f:
        pickle.dump(depth_gt, f)

INIT_ARGS = {
    "top_down_path": top_down_path,
    "btm_up_path": btm_up_path
}

ARGS = {
    "pts_out_path": pts_out_path,
    "dep_out_path": dep_out_path,
    "gt_dep_path": gt_dep_path
}

METHOD = "match"
Setup block drafted by
a language model, class construct-then-call prompt, model not recorded
Imports
numpy, tqdm — import names as written, not package names to install; a name the repository itself provides is not a package
Returned
type NoneType — a NoneType output has no array shape of its own; none is stated — the instrument coerces any output to an array; for a NoneType 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: torch and numpy seeded to 0 by the harness immediately before the class was constructed from INIT_ARGS (fixture_check.py)
procrustes Ran
ran, honoured the contract we drafted
3dpose/3D-Multi-Person-Pose/lib/posematcher.py
code served (permissive licence) · get_code("8534a20a6a9abd9a")
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

# Create small numpy arrays for predicted and target
# The function expects 2D arrays that will be transposed
# Shape: (n_points, n_features)
np.random.seed(0)

predicted = np.random.randn(4, 3).astype(np.float32)
target = np.random.randn(4, 3).astype(np.float32)

ARGS = {
    "predicted": predicted,
    "target": target,
}
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

Also present in other papers' code

Of 2 samples lifted from this paper's linked repositories, 2 carry a function identity and 0 do not. Of the 2 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 2 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

ized heatmaps, allowing the network to be more robust in handling scale variations. Finally, the estimated 3D poses from the top-down and bottom-up networks are fed into our integration network for final 3D poses. Besides the integration of top-down and bottom-up networks, unlike existing pose discriminators that are designed solely for a single person, and consequently cannot assess natural interperson interactions, we propose a two-person pose discriminator that enforces natural two-person interactions. Lastly, we also apply a semi-supervised method to overcome the 3D ground-truth data scarcity. Quantitative and qualitative evaluations show the effectiveness of the proposed method. Our code is available publicly. 1

For agents

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

get_harvested_code_for_paper("2104.01797")
get_code_for_paper("2104.01797")
have("2104.01797")

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

curl https://syntology.ai/api/ran/2104.01797.json

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

[![Syntology run record](https://syntology.ai/api/ran/2104.01797.svg)](https://syntology.ai/paper/2104.01797)

Connect an agent — have() is free.