SYNTOLOGY HomeExplorerAtlasCodeMethodologyAboutDevelopersFeedPricing
Paper · 2304.03199 · 2023

Face Animation with an Attribute-Guided Diffusion Model

arXiv · PDF · Open in the Atlas

Code that ran

12 of 12 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; 5 failures of our instrument; 0 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/2304.03199.json · badge: /api/ran/2304.03199.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
zengbohan0217/fadm canonical (classifier label)
classifier basis not recorded on the link
12 ran · 0 no recorded run
FunctionStatusWhere it lives
crop_clip Ran
ran; the contract check could not run (RAISES, DEP_MISSING or TIMEOUT)
zengbohan0217/fadm/augmentation.py
code served (permissive licence) · get_code("bc39b8ed5fefa7eb")
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 4, 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
import PIL
from PIL import Image

# Create a small numpy array-based clip
np_clip = [
    np.random.randint(0, 256, (100, 120, 3), dtype=np.uint8),
    np.random.randint(0, 256, (100, 120, 3), dtype=np.uint8),
]

# Create a small PIL Image-based clip
pil_clip = [
    Image.new('RGB', (120, 100), color=(73, 109, 137)),
    Image.new('RGB', (120, 100), color=(200, 150, 100)),
]

# Test with numpy arrays
ARGS = {
    'clip': np_clip,
    'min_h': 10,
    'min_w': 15,
    'h': 50,
    'w': 60,
}
Setup block drafted by
a language model, Claude Haiku 4.5, prompt python_fixture_v1, through the batch API
Imports
PIL, numpy — import names as written, not package names to install; a name the repository itself provides is not a package
Returned
type list, ndim 4, 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
get_dataset_distributed_ Ran
ran on a synthesized input
zengbohan0217/fadm/datasets.py
code served (permissive licence) · get_code("71ca8b5b1b0cf407")
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 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 torch
from torch.utils.data import Dataset

torch.manual_seed(0)

class SimpleDataset(Dataset):
    """Minimal dataset that returns (x, y) tuples"""
    def __init__(self, size=8):
        self.size = size
    
    def __len__(self):
        return self.size
    
    def __getitem__(self, idx):
        x = torch.randn(4)
        y = torch.randint(0, 2, (1,)).item()
        return x, y

_dataset = SimpleDataset(size=8)
world_size = 2
rank = 0
batch_size = 2

ARGS = {
    '_dataset': _dataset,
    'world_size': world_size,
    'rank': rank,
    'batch_size': batch_size
}
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 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
get_rotation_matrix Ran
ran; our contract draft was wrong, not the code
zengbohan0217/fadm/modules/Facevid/keypoint_transform.py
pointer only (licence: Apache-2.0); the stored copy get_code serves, from liluoqaq/THFEM, is pointer only (licence: NONE) · get_code("6110b3ed0b5cc79e")
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 liluoqaq/THFEM, 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.

headpose_pred_to_degree Ran
ran; our contract draft was wrong, not the code
zengbohan0217/fadm/modules/Facevid/keypoint_transform.py
pointer only (licence: Apache-2.0); the stored copy get_code serves, from liluoqaq/THFEM, is pointer only (licence: NONE) · get_code("28221b535631c490")
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 liluoqaq/THFEM, 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.

keypoint_transformation Ran
ran on a synthesized input
zengbohan0217/fadm/modules/Facevid/keypoint_transform.py
code served (permissive licence) · get_code("01fbc83284fd37de")
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 dict. 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 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.functional as F

torch.manual_seed(0)

# kp_canonical is a dict with 'value' and 'jacobian' keys
# 'value' has shape (batch, num_keypoints, 3) for 3D coordinates
# 'jacobian' has shape (batch, num_keypoints, 3, something) for derivatives
kp_canonical = {
    'value': torch.randn(2, 4, 3),  # batch=2, num_keypoints=4, coords=3
    'jacobian': torch.randn(2, 4, 3, 5)  # batch=2, num_keypoints=4, coords=3, jacobian_dim=5
}

# he is a dict with head pose predictions and transformations
# yaw, pitch, roll are softmax predictions over 66 bins
# t is translation (batch, 3)
# exp is expression coefficients (batch, num_exp_params)
he = {
    'yaw': torch.randn(2, 66),  # batch=2, 66 bins for softmax
    'pitch': torch.randn(2, 66),
    'roll': torch.randn(2, 66),
    't': torch.randn(2, 3),  # translation vector
    'exp': torch.randn(2, 12)  # expression parameters that will be reshaped to (batch, num_keypoints, 3)
}

# estimate_jacobian: bool flag
estimate_jacobian = True

# free_view: bool flag
free_view = False

# yaw, pitch, roll: optional angle overrides (only used when free_view=True)
yaw = None
pitch = None
roll = None

ARGS = {
    'kp_canonical': kp_canonical,
    'he': he,
    'estimate_jacobian': estimate_jacobian,
    'free_view': free_view,
    'yaw': yaw,
    'pitch': pitch,
    'roll': roll
}
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 dict — a dict output has no array shape of its own; none is stated — the instrument coerces any output to an array; for a dict 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
kp2gaussian Ran
ran; the contract check could not run (RAISES, DEP_MISSING or TIMEOUT)
zengbohan0217/fadm/modules/FOMM/util.py
code served (permissive licence) · get_code("e136efdcba23139b")
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 (2, 4, 8, 6) 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)

# Construct valid arguments for kp2gaussian function
# The function expects:
# - kp: a dict with 'value' key containing a tensor of keypoint coordinates
# - spatial_size: tuple of (height, width)
# - kp_variance: float for gaussian variance

# Looking at the function body:
# - kp['value'] is accessed and has a .type() method (it's a tensor)
# - mean.shape is used to determine leading dimensions
# - The function expects mean to have shape (..., 2) where ... are batch dimensions
# - spatial_size should be a tuple of (h, w)
# - kp_variance is used in division, so it's a scalar

# Create a keypoint dict with a tensor value
# Shape: (batch=2, num_keypoints=4, coordinates=2)
kp_value = torch.randn(2, 4, 2)

kp = {
    'value': kp_value
}

# Spatial size as (height, width)
spatial_size = (8, 6)

# Variance for gaussian
kp_variance = 0.01

ARGS = {
    'kp': kp,
    'spatial_size': spatial_size,
    'kp_variance': kp_variance
}
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 4, dtype family float, shape (2, 4, 8, 6) — shape (2, 4, 8, 6) 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
kp2gaussian Ran
ran on a synthesized input
zengbohan0217/fadm/modules/Facevid/util.py
code served (permissive licence) · get_code("db70096366874e50")
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 5, dtype family float, shape (2, 4, 5, 6, 7) 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)

# Analyze what kp2gaussian needs:
# - kp: a dict with a 'value' key containing a tensor
# - spatial_size: a tuple of (d, h, w) dimensions
# - kp_variance: a scalar for variance

# The function uses:
# - kp['value']: tensor that gets .type() called on it
# - mean.shape to get dimensions
# - The mean tensor is expected to have leading batch dimensions, then 3 coordinates

# Create a keypoint dict with a value tensor
# Shape: (batch=2, num_keypoints=4, coordinates=3)
kp_value = torch.randn(2, 4, 3)
kp = {'value': kp_value}

# Spatial size as (depth, height, width)
spatial_size = (5, 6, 7)

# Variance for the gaussian
kp_variance = 0.1

ARGS = {
    'kp': kp,
    'spatial_size': spatial_size,
    'kp_variance': kp_variance
}
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 5, dtype family float, shape (2, 4, 5, 6, 7) — shape (2, 4, 5, 6, 7) 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
make_coordinate_grid Ran
ran; the contract check could not run (RAISES, DEP_MISSING or TIMEOUT)
zengbohan0217/fadm/modules/FOMM/util.py
code served (permissive licence) · get_code("fe3cbec3030f03b9")
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 (4, 3, 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

torch.manual_seed(0)

# spatial_size: tuple of (height, width)
spatial_size = (4, 3)

# type: torch dtype
type_arg = torch.float32

ARGS = {
    "spatial_size": spatial_size,
    "type": type_arg,
}
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 (4, 3, 2) — shape (4, 3, 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
make_coordinate_grid Ran
ran on a synthesized input
zengbohan0217/fadm/modules/Facevid/util.py
code served (permissive licence) · get_code("ea86605026618a21")
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 (2, 3, 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)

spatial_size = (2, 3, 4)
type_arg = torch.float32

ARGS = {
    "spatial_size": spatial_size,
    "type": type_arg
}
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 4, dtype family float, shape (2, 3, 4, 3) — shape (2, 3, 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
make_coordinate_grid_2d Ran
ran on a synthesized input
zengbohan0217/fadm/modules/Facevid/util.py
code served (permissive licence) · get_code("a19a2d7e79c6fcbd")
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 (4, 3, 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

torch.manual_seed(0)

# The function takes spatial_size (a tuple of height and width)
# and type (a torch dtype)

ARGS = {
    "spatial_size": (4, 3),
    "type": torch.float32
}
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 (4, 3, 2) — shape (4, 3, 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
read_video Ran
ran on a synthesized input
zengbohan0217/fadm/frames_dataset.py
code served (permissive licence) · get_code("47f88e1a3fa42f72")
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 (3, 16, 20, 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 os
import tempfile
import numpy as np
from skimage import io, img_as_float32
from skimage.color import gray2rgb
from PIL import Image

# Create a temporary directory with test frames
temp_dir = tempfile.mkdtemp()

# Create a few simple test images as frames
for i in range(3):
    img_array = np.random.randint(0, 256, (16, 20, 3), dtype=np.uint8)
    img_path = os.path.join(temp_dir, f"frame_{i:03d}.png")
    Image.fromarray(img_array).save(img_path)

# Define frame_shape for the test
frame_shape = (16, 20, 3)

ARGS = {
    "name": temp_dir,
    "frame_shape": frame_shape
}
Setup block drafted by
a language model, Claude Haiku 4.5, prompt python_fixture_v1, through the batch API
Imports
PIL, imageio, numpy, skimage — 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 (3, 16, 20, 3) — shape (3, 16, 20, 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
resize_clip Ran
ran on a synthesized input
zengbohan0217/fadm/augmentation.py
code served (permissive licence) · get_code("6157582222ab30fa")
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 4, 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 PIL
from PIL import Image

# Create a small numpy array clip (list of numpy arrays)
# Shape: (height=10, width=8, channels=3) - distinct dimensions
numpy_frame_1 = np.random.randint(0, 255, (10, 8, 3), dtype=np.uint8)
numpy_frame_2 = np.random.randint(0, 255, (10, 8, 3), dtype=np.uint8)
numpy_clip = [numpy_frame_1, numpy_frame_2]

# Create a PIL Image clip (list of PIL Images)
pil_frame_1 = Image.fromarray(numpy_frame_1)
pil_frame_2 = Image.fromarray(numpy_frame_2)
pil_clip = [pil_frame_1, pil_frame_2]

ARGS = {
    'clip': numpy_clip,
    'size': 256,
    'interpolation': 'bilinear'
}
Setup block drafted by
a language model, Claude Haiku 4.5, prompt python_fixture_v1, through the batch API
Imports
PIL, numpy, skimage — import names as written, not package names to install; a name the repository itself provides is not a package
Returned
type list, ndim 4, 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

Also present in other papers' code

Of 12 samples lifted from this paper's linked repositories, 12 carry a function identity and 0 do not. Of the 12 distinct functions with an identity, 7 are also present in other papers whose code we hold (5 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
crop_clip 13 other papers
4 dated earlier, 9 later
2112.02815 (2021-12)
Make It Move: Controllable Image-to-Video Generation with Text Descriptions
2209.10340 (2022-09)
FNeVR: Neural Volume Rendering for Face Animation
2210.00788 (2022-10)
Towards a Unified View on Visual Parameter-Efficient Transfer Learning
and 10 more not named here
make_coordinate_grid 7 other papers
5 dated earlier, 2 later · 4 only as an assembly brick
2003.00196 (2020-03)
First Order Motion Model for Image Animation
2104.11280 (2021-04)
Motion Representations for Articulated Animation
2203.14367 (2022-03)
Thin-Plate Spline Motion Model for Image Animation
and 4 more not named here
kp2gaussian 3 other papers
3 dated earlier · 1 only as an assembly brick
2003.00196 (2020-03)
First Order Motion Model for Image Animation
2204.05018 (2022-04)
Structure-Aware Motion Transfer with Deformable Anchor Model
2209.10340 (2022-09)
FNeVR: Neural Volume Rendering for Face Animation
get_rotation_matrix 1 other paper
1 later
2601.12876 (2026-01)
Exploring Talking Head Models With Adjacent Frame Prior for Speech-Preserving Facial Expression Manipulation
headpose_pred_to_degree 1 other paper
1 later
2601.12876 (2026-01)
Exploring Talking Head Models With Adjacent Frame Prior for Speech-Preserving Facial Expression Manipulation
make_coordinate_grid 1 other paper
1 dated earlier
2209.10340 (2022-09)
FNeVR: Neural Volume Rendering for Face Animation
resize_clip 1 other paper
1 dated earlier
2209.10340 (2022-09)
FNeVR: Neural Volume Rendering for Face Animation

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

Face animation has achieved much progress in computer vision. However, prevailing GAN-based methods suffer from unnatural distortions and artifacts due to sophisticated motion deformation. In this paper, we propose a Face Animation framework with an attribute-guided Diffusion Model (FADM), which is the first work to exploit the superior modeling capacity of diffusion models for photo-realistic talking-head generation. To mitigate the uncontrollable synthesis effect of the diffusion model, we design an Attribute-Guided Conditioning Network (AGCN) to adaptively combine the coarse animation features and 3D face reconstruction results, which can incorporate appearance and motion conditions into the diffusion process. These specific designs help FADM rectify unnatural artifacts and distortions, and also enrich high-fidelity facial details through iterative diffusion refinements with accurate animation attributes. FADM can flexibly and effectively improve existing animation videos. Extensive experiments on widely used talking-head benchmarks validate the effectiveness of FADM over prior arts.

For agents

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

get_harvested_code_for_paper("2304.03199")
get_code_for_paper("2304.03199")
have("2304.03199")

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

curl https://syntology.ai/api/ran/2304.03199.json

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

[![Syntology run record](https://syntology.ai/api/ran/2304.03199.svg)](https://syntology.ai/paper/2304.03199)

Connect an agent — have() is free.