SYNTOLOGY HomeExplorerAtlasCodeMethodologyAboutDevelopersFeedPricing
Paper · 1904.01189 · 2019

Semantics-Guided Neural Networks for Efficient Skeleton-Based Human Action Recognition

arXiv · PDF · Open in the Atlas

Code that ran

11 of 13 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 8 have no contract verdict; 3 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/1904.01189.json · badge: /api/ran/1904.01189.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
microsoft/SGN canonical (classifier label)
classifier basis: classifier_2025
3 ran · 0 no recorded run
urw7rs/torch_skeleton community (classifier label)
classifier basis not recorded on the link
8 ran · 2 no recorded run
FunctionStatusWhere it lives
denoising_by_length Ran
ran; our contract draft was wrong, not the code
microsoft/SGN/data/ntu/get_raw_denoised_data.py
code served (permissive licence) · get_code("d63c7bd0fe3e1881")
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 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 logging

# Set up logging to avoid warnings
logging.basicConfig(level=logging.INFO)

# Define ARGS with valid arguments for denoising_by_length
ARGS = {
    'ske_name': 'skeleton_001',
    'bodies_data': {
        'body_1': {
            'interval': [0, 1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12],  # length 13, > threshold
            'motion': 0.75
        },
        'body_2': {
            'interval': [0, 1, 2],  # length 3, <= threshold (will be filtered)
            'motion': 0.25
        },
        'body_3': {
            'interval': [0, 1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11],  # length 12, > threshold
            'motion': 0.50
        },
        'body_4': {
            'interval': [0],  # length 1, <= threshold (will be filtered)
            'motion': 0.10
        }
    }
}
Setup block drafted by
a language model, Claude Haiku 4.5, prompt python_fixture_v1
Imports
none outside the standard library — 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
denoising_by_spread Ran
ran; our contract draft was wrong, not the code
microsoft/SGN/data/ntu/get_raw_denoised_data.py
code served (permissive licence) · get_code("a521d7380bf90c08")
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 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 numpy as np
import logging

# Set up logging to avoid errors
logging.basicConfig(level=logging.INFO)

# Create test data for bodies_data
# The function expects:
# - bodies_data: dict with bodyID keys
# - each body_data has: 'joints' (reshaped to -1, 25, 3), 'interval', 'motion'

num_frames = 10
num_joints = 25
joint_dim = 3

# Create joints data: (num_frames * 25 * 3,) flattened array
joints_data = np.random.rand(num_frames * num_joints * joint_dim).astype(np.float32)

# Create bodies_data dict with 2 bodyIDs
bodies_data = {
    'body_0': {
        'joints': joints_data.copy(),
        'interval': list(range(num_frames)),
        'motion': 0.5
    },
    'body_1': {
        'joints': joints_data.copy() + np.random.rand(num_frames * num_joints * joint_dim).astype(np.float32) * 0.1,
        'interval': list(range(num_frames)),
        'motion': 0.3
    }
}

ske_name = 'test_skeleton_001'

ARGS = {
    'ske_name': ske_name,
    'bodies_data': bodies_data
}
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 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
downloaded Ran
ran on a synthesized input
urw7rs/torch_skeleton/torch_skeleton/utils.py
code served (permissive licence) · get_code("6b0cdc54f0be85d8")
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 os
import tempfile

# Create a temporary directory and file for testing
temp_dir = tempfile.mkdtemp()
test_file = os.path.join(temp_dir, "test_file.txt")

# Create the file so it exists
with open(test_file, 'w') as f:
    f.write("test content")

ARGS = {
    "path": test_file
}
Setup block drafted by
a language model, Claude Haiku 4.5, prompt python_fixture_v1, through the batch API
Imports
none outside the standard library — 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
filter_split Ran
ran on a synthesized input
urw7rs/torch_skeleton/torch_skeleton/datasets/ucla.py
code served (permissive licence) · get_code("1fca1af804c2a135")
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 os.path as osp

# The function filter_split takes:
# 1. paths: a list of path strings that will be processed by get_camera()
# 2. split: a string that is either 'train' or 'test'

# get_camera() expects paths with format: "..._{camera_id}..." where camera_id is like "c0", "c1", "c2", "c3"
# It extracts the basename, splits by '.', takes first part, splits by '_', takes index 3, and removes first char

# Let's construct realistic file paths that will work with get_camera()
paths = [
    "/data/subject_01_seq_01_c0_frame_001.jpg",  # camera 0 - in train split
    "/data/subject_01_seq_01_c1_frame_001.jpg",  # camera 1 - in train split
    "/data/subject_01_seq_01_c2_frame_001.jpg",  # camera 2 - in train split
    "/data/subject_01_seq_01_c3_frame_001.jpg",  # camera 3 - in test split (camera_id == 3)
    "/data/subject_01_seq_02_c0_frame_001.jpg",  # camera 0 - in train split
]

split = 'train'

ARGS = {
    'paths': paths,
    'split': split,
}
Setup block drafted by
a language model, Claude Haiku 4.5, prompt python_fixture_v1, through the batch API
Imports
none outside the standard library — 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
get_camera Ran
ran on a synthesized input
urw7rs/torch_skeleton/torch_skeleton/datasets/ucla.py
code served (permissive licence) · get_code("000b53fddf4c7be7")
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 int, ndim 0, dtype family int, 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 os.path as osp
import tempfile

# Create a temporary file with a name matching the expected pattern
# The function extracts the camera ID from the 4th underscore-separated component
# Format: something_something_something_cX.ext where X is the camera number
temp_dir = tempfile.gettempdir()
test_path = osp.join(temp_dir, "file_name_part_c5.txt")

ARGS = {
    "path": test_path
}
Setup block drafted by
a language model, Claude Haiku 4.5, prompt python_fixture_v1, through the batch API
Imports
none outside the standard library — import names as written, not package names to install; a name the repository itself provides is not a package
Returned
type int, ndim 0, dtype family int, 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
get_filename Ran
ran on a synthesized input
urw7rs/torch_skeleton/torch_skeleton/datasets/ntu.py
code served (permissive licence) · get_code("aeefc2fb74373e0a")
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 str. 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 os.path as osp

# The function takes a file_path string and extracts the filename without extension
# using osp.basename() and split('.')

ARGS = {
    'file_path': '/path/to/example_file.txt'
}
Setup block drafted by
a language model, Claude Haiku 4.5, prompt python_fixture_v1, through the batch API
Imports
none outside the standard library — import names as written, not package names to install; a name the repository itself provides is not a package
Returned
type str — a str output has no array shape of its own; none is stated — the instrument coerces any output to an array; for a str 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_indices Ran
ran on a synthesized input
urw7rs/torch_skeleton/torch_skeleton/transforms/functional.py
code served (permissive licence) · get_code("756a7b2dc4c4f0c7")
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 2, 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

# Create a small 4D numpy array
# The function calls x.sum(axis=(2,3)), so we need at least 4 dimensions
# Shape: (2, 3, 4, 5) - batch=2, channels=3, height=4, width=5
x = np.random.randint(0, 2, size=(2, 3, 4, 5)).astype(np.float32)

# Ensure at least some non-zero sums along axes (2,3) for valid output
x[0, 0, :, :] = 1.0  # Make sure at least one position has non-zero sum

ARGS = {
    "x": x,
    "axis": (2, 3)
}
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 tuple, ndim 2, 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
get_mask Ran
ran on a synthesized input
urw7rs/torch_skeleton/torch_skeleton/transforms/functional.py
code served (permissive licence) · get_code("0467f83477d7c817")
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 4, dtype family float, shape (2, 3, 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

# Create a small 4D numpy array (batch=2, channels=3, height=4, width=5)
# The function uses x.sum(axis=(2, 3)) so we need at least 4 dimensions
x = np.random.randn(2, 3, 4, 5)

# Ensure at least one slice has non-zero sum along axes 2,3 for interesting behavior
x[0, 0, :, :] = np.ones((4, 5))  # This will have non-zero sum

ARGS = {
    "x": x
}
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 4, dtype family float, shape (2, 3, 4, 5) — shape (2, 3, 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_valid_frames_by_spread Ran
ran; the contract check could not run (RAISES, DEP_MISSING or TIMEOUT)
microsoft/SGN/data/ntu/get_raw_denoised_data.py
code served (permissive licence) · get_code("7bdf5bd91eaabb7e")
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 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
import numpy as np

torch.manual_seed(0)

# The function expects 'points' to be array-like with shape (num_frames, num_joints, 2)
# where the last dimension contains [x, y] coordinates
# It accesses points[i, :, 0] for x coordinates and points[i, :, 1] for y coordinates

# Create a small tensor with shape (num_frames=2, num_joints=4, coordinates=2)
points = torch.randn(2, 4, 2)

ARGS = {
    "points": points
}
Setup block drafted by
a language model, Claude Haiku 4.5, prompt python_fixture_v1
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, 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
listdir Ran
ran on a synthesized input
urw7rs/torch_skeleton/torch_skeleton/utils.py
code served (permissive licence) · get_code("c4b7f9203fdb9f3d")
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 os
import tempfile
import shutil

# Create a temporary directory structure for testing
test_dir = tempfile.mkdtemp()

# Create subdirectories and files
os.makedirs(os.path.join(test_dir, "subdir1"), exist_ok=True)
os.makedirs(os.path.join(test_dir, "subdir2"), exist_ok=True)
os.makedirs(os.path.join(test_dir, "subdir1", "nested"), exist_ok=True)

# Create test files with various extensions
open(os.path.join(test_dir, "file1.txt"), "w").close()
open(os.path.join(test_dir, "file2.py"), "w").close()
open(os.path.join(test_dir, "subdir1", "file3.txt"), "w").close()
open(os.path.join(test_dir, "subdir1", "file4.md"), "w").close()
open(os.path.join(test_dir, "subdir1", "nested", "file5.txt"), "w").close()
open(os.path.join(test_dir, "subdir2", "file6.txt"), "w").close()

ARGS = {
    "root": test_dir,
    "ext": ".txt"
}
Setup block drafted by
a language model, Claude Haiku 4.5, prompt python_fixture_v1, through the batch API
Imports
none outside the standard library — 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
to_numpy Ran
ran on a synthesized input
urw7rs/torch_skeleton/torch_skeleton/datasets/ntu.py
code served (permissive licence) · get_code("847a989802c341f5")
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 4, dtype family float, shape (2, 3, 25, 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
from typing import Any, List, Dict

# Construct a valid skeleton_sequence dictionary matching the function's expectations
skeleton_sequence = {
    'numFrames': 3,  # 3 frames
    'frames': [
        {
            'numBodies': 2,
            'bodies': [
                {
                    'joints': [
                        {
                            'cameraX': 0.1 * i,
                            'cameraY': 0.2 * i,
                            'cameraZ': 0.3 * i
                        }
                        for i in range(1, 26)  # 25 joints
                    ]
                },
                {
                    'joints': [
                        {
                            'cameraX': 0.15 * i,
                            'cameraY': 0.25 * i,
                            'cameraZ': 0.35 * i
                        }
                        for i in range(1, 26)  # 25 joints
                    ]
                }
            ]
        },
        {
            'numBodies': 1,
            'bodies': [
                {
                    'joints': [
                        {
                            'cameraX': 0.05 * i,
                            'cameraY': 0.1 * i,
                            'cameraZ': 0.15 * i
                        }
                        for i in range(1, 26)  # 25 joints
                    ]
                }
            ]
        },
        {
            'numBodies': 2,
            'bodies': [
                {
                    'joints': [
                        {
                            'cameraX': 0.2 * i,
                            'cameraY': 0.3 * i,
                            'cameraZ': 0.4 * i
                        }
                        for i in range(1, 26)  # 25 joints
                    ]
                },
                {
                    'joints': [
                        {
                            'cameraX': 0.12 * i,
                            'cameraY': 0.22 * i,
                            'cameraZ': 0.32 * i
                        }
                        for i in range(1, 26)  # 25 joints
                    ]
                }
            ]
        }
    ]
}

joint_keys = ['cameraX', 'cameraY', 'cameraZ']

ARGS = {
    'skeleton_sequence': skeleton_sequence,
    'joint_keys': joint_keys
}
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 4, dtype family float, shape (2, 3, 25, 3) — shape (2, 3, 25, 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
loads Not yet run
unverified (harvested, no recorded run)
urw7rs/torch_skeleton/torch_skeleton/datasets/ntu.py
code served (permissive licence) · get_code("d780706a227ca858")
How it ran

no invocation recorded.

pad_bodies Not yet run
unverified (harvested, no recorded run)
urw7rs/torch_skeleton/torch_skeleton/transforms/functional.py
code served (permissive licence) · get_code("42c7e6ad59d7eafd")
How it ran

no invocation recorded.

Also present in other papers' code

Of 13 samples lifted from this paper's linked repositories, 13 carry a function identity and 0 do not. Of the 13 distinct functions with an identity, 3 are also present in other papers whose code we hold (0 with at least one paper dated earlier, 0 the same month, 3 later) and 10 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
denoising_by_length 1 other paper
1 later
2201.02849 (2022-01)
Spatio-Temporal Tuples Transformer for Skeleton-Based Action Recognition
denoising_by_spread 1 other paper
1 later
2201.02849 (2022-01)
Spatio-Temporal Tuples Transformer for Skeleton-Based Action Recognition
get_valid_frames_by_spread 1 other paper
1 later
2201.02849 (2022-01)
Spatio-Temporal Tuples Transformer for Skeleton-Based Action Recognition

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

Skeleton-based human action recognition has attracted great interest thanks to the easy accessibility of the human skeleton data. Recently, there is a trend of using very deep feedforward neural networks to model the 3D coordinates of joints without considering the computational efficiency. In this paper, we propose a simple yet effective semantics-guided neural network (SGN) for skeleton-based action recognition. We explicitly introduce the high level semantics of joints (joint type and frame index) into the network to enhance the feature representation capability. In addition, we exploit the relationship of joints hierarchically through two modules, i.e., a joint-level module for modeling the correlations of joints in the same frame and a framelevel module for modeling the dependencies of frames by taking the joints in the same frame as a whole. A strong baseline is proposed to facilitate the study of this field. With an order of magnitude smaller model size than most previous works, SGN achieves the state-of-the-art performance on the NTU60, NTU120, and SYSU datasets. The source code is available at https://github.com/microsoft/SGN.

For agents

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

get_harvested_code_for_paper("1904.01189")
get_code_for_paper("1904.01189")
have("1904.01189")

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

curl https://syntology.ai/api/ran/1904.01189.json

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

[![Syntology run record](https://syntology.ai/api/ran/1904.01189.svg)](https://syntology.ai/paper/1904.01189)

Connect an agent — have() is free.