SYNTOLOGY HomeExplorerAtlasCodeMethodologyAboutDevelopersFeedPricing
Paper · 1703.07332 · 2017

How far are we from solving the 2D & 3D Face Alignment problem? (and a dataset of 230,000 3D facial landmarks)

arXiv · PDF · Open in the Atlas

Code that ran

14 of 15 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 7 have no contract verdict; 7 failures of our instrument; 1 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/1703.07332.json · badge: /api/ran/1703.07332.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
1adrianb/face-alignment canonical (classifier label)
classifier basis: census
12 ran · 1 no recorded run
GuohongLi/face-alignment-pytorch community (classifier label)
classifier basis not recorded on the link
2 ran · 0 no recorded run
FunctionStatusWhere it lives
conv3x3 Ran
ran; our contract draft was wrong, not the code
GuohongLi/face-alignment-pytorch/models/resnet.py
pointer only (licence: BSD-3-Clause); the stored copy get_code serves, from lixiaotong97/DSU, is pointer only (licence: NONE) · get_code("fac5364e2f53c6db")
How it ran

The body of this sample is withheld: this paper's link allows the body inline (BSD-3-Clause), but get_code serves the stored copy, first harvested from lixiaotong97/DSU, 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.

conv3x3 Ran
ran; our contract draft was wrong, not the code
1adrianb/face-alignment/face_alignment/models/fan.py
pointer only (licence: BSD-3-Clause); the stored copy get_code serves, from xiexh20/CHORE, is pointer only (licence: NONE) · get_code("5baaa8c1b148ef70")
How it ran

The body of this sample is withheld: this paper's link allows the body inline (BSD-3-Clause), but get_code serves the stored copy, first harvested from xiexh20/CHORE, 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.

conv_bn Ran
ran; our contract draft was wrong, not the code
1adrianb/face-alignment/face_alignment/detection/retinaface/mobilenet.py
code served (permissive licence) · get_code("f5414e78ff7a815f")
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 Sequential. 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)

# The function conv_bn takes inp, oup, stride, and leaky parameters
# inp: input channels (int)
# oup: output channels (int)
# stride: stride for Conv2d (int, default=1)
# leaky: negative_slope for LeakyReLU (float, default=0)

ARGS = {
    'inp': 3,
    'oup': 16,
    'stride': 1,
    'leaky': 0.1
}
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 Sequential — a Sequential output has no array shape of its own; none is stated — the instrument coerces any output to an array; for a Sequential output the recorded ndim 1 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
conv_bn1x1 Ran
ran on a synthesized input
1adrianb/face-alignment/face_alignment/detection/retinaface/net_retinaface.py
code served (permissive licence) · get_code("e8b6601f1ef32d21")
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 Sequential. 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)

ARGS = {
    "inp": 3,
    "oup": 16,
    "stride": 1,
    "leaky": 0.1
}
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 Sequential — a Sequential output has no array shape of its own; none is stated — the instrument coerces any output to an array; for a Sequential output the recorded ndim 1 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
conv_bn_no_relu Ran
ran; our contract draft was wrong, not the code
1adrianb/face-alignment/face_alignment/detection/retinaface/net_retinaface.py
code served (permissive licence) · get_code("aacb8f654f54d534")
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 Sequential. 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)

ARGS = {
    "inp": 3,
    "oup": 16,
    "stride": 1
}
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 Sequential — a Sequential output has no array shape of its own; none is stated — the instrument coerces any output to an array; for a Sequential output the recorded ndim 1 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
conv_dw Ran
ran on a synthesized input
1adrianb/face-alignment/face_alignment/detection/retinaface/mobilenet.py
code served (permissive licence) · get_code("e2e512278cdc740a")
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 Sequential. 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)

ARGS = {
    "inp": 3,
    "oup": 16,
    "stride": 1,
    "leaky": 0.1
}
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 Sequential — a Sequential output has no array shape of its own; none is stated — the instrument coerces any output to an array; for a Sequential output the recorded ndim 1 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
crop Ran
ran on a synthesized input
1adrianb/face-alignment/face_alignment/utils.py
code served (permissive licence) · get_code("e6d17ffcdd378aab")
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 int, shape (256, 256, 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 numpy as np
import cv2

torch.manual_seed(0)

# Create a small test image (3D RGB image)
image = np.random.randint(0, 256, (64, 64, 3), dtype=np.uint8)

# Create center point as numpy array
center = np.array([32.0, 32.0])

# Define scale
scale = 0.5

# Define resolution
resolution = 256.0

ARGS = {
    "image": image,
    "center": center,
    "scale": scale,
    "resolution": resolution,
}
Setup block drafted by
a language model, Claude Haiku 4.5, prompt python_fixture_v1, through the batch API
Imports
cv2, numpy, torch — 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 int, shape (256, 256, 3) — shape (256, 256, 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
draw_gaussian Ran
ran on a synthesized input
1adrianb/face-alignment/face_alignment/utils.py
code served (permissive licence) · get_code("e06cb5e221fe3fb8")
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 (64, 64) 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 math
import numpy as np

# First, let's build the gaussian kernel that the function uses
def _gaussian(size=3, sigma=0.25, amplitude=1, normalize=False, width=None, height=None, sigma_horz=None, sigma_vert=None, mean_horz=0.5, mean_vert=0.5):
    if width is None:
        width = size
    if height is None:
        height = size
    if sigma_horz is None:
        sigma_horz = sigma
    if sigma_vert is None:
        sigma_vert = sigma
    center_x = mean_horz * width + 0.5
    center_y = mean_vert * height + 0.5
    gauss = np.empty((height, width), dtype=np.float32)
    for i in range(height):
        for j in range(width):
            gauss[i][j] = amplitude * math.exp(-(math.pow((j + 1 - center_x) / (sigma_horz * width), 2) / 2.0 + math.pow((i + 1 - center_y) / (sigma_vert * height), 2) / 2.0))
    if normalize:
        gauss = gauss / np.sum(gauss)
    return gauss

_GAUSS_KERNEL = _gaussian(size=13)

# Create a sample image - needs to be a 2D numpy array that supports indexing and comparison
# The function uses: image.shape[0], image.shape[1], and modifies image with slicing
image = np.zeros((64, 64), dtype=np.float32)

# Create a point - needs to be indexable with [0] and [1]
# Looking at the function, point represents (x, y) coordinates
point = [32.5, 32.5]

# Create sigma - a scalar value for the Gaussian spread
sigma = 2.0

ARGS = {
    "image": image,
    "point": point,
    "sigma": sigma,
}
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 (64, 64) — shape (64, 64) 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
flip_detect Ran
ran on a synthesized input
1adrianb/face-alignment/face_alignment/detection/core.py
code served (permissive licence) · get_code("5bd43ede69fb8b7d")
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 (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 cv2
import numpy as np
import torch
import torch.nn as nn

torch.manual_seed(0)

class DummyNet(nn.Module):
    def __init__(self):
        super().__init__()
        self.linear = nn.Linear(10, 5)
    
    def forward(self, x):
        return self.linear(x)

def dummy_detect_fn(net, img, device):
    """Dummy detection function that returns bboxes with 5 columns (x1, y1, x2, y2, confidence)"""
    # Return 2 bounding boxes with format: [x1, y1, x2, y2, confidence]
    bboxes = np.array([
        [10.0, 20.0, 50.0, 80.0, 0.9],
        [60.0, 30.0, 120.0, 100.0, 0.85]
    ], dtype=np.float32)
    return bboxes

# Create a dummy image (height=200, width=300, 3 channels for BGR)
dummy_img = np.random.randint(0, 255, (200, 300, 3), dtype=np.uint8)

# Create network and device
net = DummyNet()
device = 'cpu'

ARGS = {
    'net': net,
    'img': dummy_img,
    'device': device,
    'detect_fn': dummy_detect_fn
}
Setup block drafted by
a language model, Claude Haiku 4.5, prompt python_fixture_v1, through the batch API
Imports
cv2, numpy, torch — 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 (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
intersect Ran
ran; our contract draft was wrong, not the code
1adrianb/face-alignment/face_alignment/detection/blazeface/net_blazeface.py
pointer only (licence: BSD-3-Clause); the stored copy get_code serves, from NVlabs/AL-MDN, is pointer only (licence: NOASSERTION) · get_code("f01c5f73515770a5")
How it ran

The body of this sample is withheld: this paper's link allows the body inline (BSD-3-Clause), but get_code serves the stored copy, first harvested from NVlabs/AL-MDN, and GitHub could not classify the licence of that repository (NOASSERTION). 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.

jaccard Ran
ran; our contract draft was wrong, not the code
1adrianb/face-alignment/face_alignment/detection/blazeface/net_blazeface.py
pointer only (licence: BSD-3-Clause); the stored copy get_code serves, from NVlabs/AL-MDN, is pointer only (licence: NOASSERTION) · get_code("86489779d4c76926")
How it ran

The body of this sample is withheld: this paper's link allows the body inline (BSD-3-Clause), but get_code serves the stored copy, first harvested from NVlabs/AL-MDN, and GitHub could not classify the licence of that repository (NOASSERTION). 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.

ldmk_loss Ran
ran; our contract draft was wrong, not the code
GuohongLi/face-alignment-pytorch/utils/loss.py
code served (permissive licence) · get_code("2ee40a765c522617")
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 expects:
# - input: tensor of shape (n, c) where c is even (number of landmarks * 2)
# - target: tensor of shape (n, c) with landmark coordinates
# - weight: optional, not used in the function body
# - size_average: boolean flag

# Based on the function body:
# - It accesses target[:, 36*2:37*2] and target[:, 45*2:46*2]
# - So c must be at least 46*2 = 92
# - input and target are compared element-wise
# - The function iterates through c//2 landmarks

n = 2  # batch size
c = 92  # 46 landmarks * 2 coordinates (minimum needed for indices 36-45)

# Create input tensor: shape (2, 92)
input_tensor = torch.randn(n, c).float()

# Create target tensor: shape (2, 92)
# Make sure landmarks 36 and 45 are different so iod is non-zero
target_tensor = torch.randn(n, c).float()

# weight is optional and not used in the function body
weight = None

# size_average is a boolean flag
size_average = True

ARGS = {
    'input': input_tensor,
    'target': target_tensor,
    'weight': weight,
    'size_average': size_average
}
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
pts_to_bb Ran
ran on a synthesized input
1adrianb/face-alignment/face_alignment/detection/core.py
code served (permissive licence) · get_code("ab72b4766d59c85e")
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 1, dtype family float, shape (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 numpy as np

# pts_to_bb expects a 2D array of points (N, 2) where N is the number of points
# and each point has x, y coordinates
pts = np.array([
    [1.0, 2.0],
    [3.0, 4.0],
    [2.5, 1.5],
    [0.5, 3.5]
])

ARGS = {
    "pts": pts
}
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 1, dtype family float, shape (4,) — shape (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
transform Ran
ran on a synthesized input
1adrianb/face-alignment/face_alignment/utils.py
code served (permissive licence) · get_code("18816f434cfff85c")
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 1, dtype family int, shape (2,) 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 numpy as np

torch.manual_seed(0)

# Construct arguments for the transform function
point = torch.tensor([50.0, 75.0])
center = np.array([100.0, 120.0])
scale = 1.5
resolution = 64.0

ARGS = {
    "point": point,
    "center": center,
    "scale": scale,
    "resolution": resolution,
    "invert": False
}
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 Tensor, ndim 1, dtype family int, shape (2,) — shape (2,) 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
overlap_similarity Not yet run
unverified (harvested, no recorded run)
1adrianb/face-alignment/face_alignment/detection/blazeface/net_blazeface.py
code served (permissive licence) · get_code("a0ca886f22744fea")
How it ran

no invocation recorded.

Also present in other papers' code

Of 15 samples lifted from this paper's linked repositories, 15 carry a function identity and 0 do not. Of the 15 distinct functions with an identity, 7 are also present in other papers whose code we hold (2 with at least one paper dated earlier, 1 the same month, 7 later) and 8 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
conv3x3 686 other papers
22 dated earlier, 664 later · 31 only as an assembly brick
1406.2661 (2014-06)
Generative Adversarial Networks
1409.0575 (2014-09)
ImageNet Large Scale Visual Recognition Challenge
1412.6572 (2014-12)
Explaining and Harnessing Adversarial Examples
and 683 more not named here
intersect 14 other papers
2 dated earlier, 1 the same month, 11 later
1409.1556 (2014-09)
Very Deep Convolutional Networks for Large-Scale Image Recognition
1512.02325 (2015-12)
SSD: Single Shot MultiBox Detector
and 11 more not named here
conv3x3 7 other papers
7 later
1807.07860 (2018-07)
Talking Face Generation by Adversarially Disentangled Audio-Visual Representation
2005.08209 (2020-05)
Learning Individual Speaking Styles for Accurate Lip to Speech Synthesis
2204.02445 (2022-04)
CHORE: Contact, Human and Object REconstruction from a single RGB image
and 4 more not named here
conv_bn 4 other papers
4 later
1905.00641 (2019-05)
RetinaFace: Single-stage Dense Face Localisation in the Wild
2203.11593 (2022-03)
Unified Negative Pair Generation toward Well-discriminative Feature Space for Face Recognition
2303.17176 (2023-03)
A View From Somewhere: Human-Centric Face Representations
and 1 more not named here
conv_bn_no_relu 4 other papers
4 later
1905.00641 (2019-05)
RetinaFace: Single-stage Dense Face Localisation in the Wild
2203.11593 (2022-03)
Unified Negative Pair Generation toward Well-discriminative Feature Space for Face Recognition
2303.17176 (2023-03)
A View From Somewhere: Human-Centric Face Representations
and 1 more not named here
jaccard 2 other papers
2 later
2006.10214 (2020-06)
MediaPipe Hands: On-device Real-time Hand Tracking
2103.16130 (2021-03)
Active Learning for Deep Object Detection via Probabilistic Modeling
overlap_similarity 1 other paper
1 later
2006.10214 (2020-06)
MediaPipe Hands: On-device Real-time Hand Tracking

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

This paper investigates how far a very deep neural network is from attaining close to saturating performance on existing 2D and 3D face alignment datasets. To this end, we make the following 5 contributions: (a) we construct, for the first time, a very strong baseline by combining a state-of-the-art architecture for landmark localization with a state-of-the-art residual block, train it on a very large yet synthetically expanded 2D facial landmark dataset and finally evaluate it on all other 2D facial landmark datasets. (b) We create a guided by 2D landmarks network which converts 2D landmark annotations to 3D and unifies all existing datasets, leading to the creation of LS3D-W, the largest and most challenging 3D facial landmark dataset to date ~230,000 images. (c) Following that, we train a neural network for 3D face alignment and evaluate it on the newly introduced LS3D-W. (d) We further look into the effect of all "traditional" factors affecting face alignment performance like large pose, initialization and resolution, and introduce a "new" one, namely the size of the network. (e) We show that both 2D and 3D face alignment networks achieve performance of remarkable accuracy which is probably close to saturating the datasets used. Training and testing code as well as the dataset can be downloaded from https://www.adrianbulat.com/face-alignment/

For agents

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

get_harvested_code_for_paper("1703.07332")
get_code_for_paper("1703.07332")
have("1703.07332")

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

curl https://syntology.ai/api/ran/1703.07332.json

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

[![Syntology run record](https://syntology.ai/api/ran/1703.07332.svg)](https://syntology.ai/paper/1703.07332)

Connect an agent — have() is free.