SYNTOLOGY HomeExplorerAtlasCodeMethodologyAboutDevelopersFeedPricing
Paper · 2203.07967 · ECCV · 2022

Intrinsic Neural Fields: Learning Functions on Manifolds

Michael Moeller, Daniel Cremers, Zorah Lähner, Lukas Koestler, Daniel Grittner

arXiv · PDF · Open in the Atlas

Code that ran

6 of 14 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 5 have no contract verdict; 0 failures of our instrument; 8 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/2203.07967.json · badge: /api/ran/2203.07967.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
tum-vision/intrinsic-neural-fields marked official in Papers with Code
Papers with Code archive
6 ran · 8 no recorded run
FunctionStatusWhere it lives
FourierFeatEnc Ran
ran on a synthesized input
tum-vision/intrinsic-neural-fields/neutex/neutex.py
code served (permissive licence) · get_code("eaec0a6151d6c87a")
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 Tensor, ndim 2, dtype family float, shape (2, 28) 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 torch.nn as nn

class FourierFeatEnc(nn.Module):
    """
    Inspired by
    https://github.com/facebookresearch/pytorch3d/blob/fc4dd80208bbcf6f834e7e1594db0e106961fd60/pytorch3d/renderer/implicit/harmonic_embedding.py#L10
    """

    def __init__(self, k, include_input=True, use_logspace=True, max_freq=None):
        super(FourierFeatEnc, self).__init__()
        if use_logspace:
            freq_bands = 2 ** torch.arange(0, k) * torch.pi
        else:
            assert max_freq is not None
            freq_bands = 2 ** torch.linspace(0, max_freq, steps=k + 1)[:-1] * torch.pi
        self.register_buffer('freq_bands', freq_bands, persistent=False)
        self.include_input = include_input

    def forward(self, x):
        embed = (x[..., None] * self.freq_bands).view(*x.size()[:-1], -1)
        if self.include_input:
            return torch.cat((embed.cos(), embed.sin(), x), dim=-1)
        return torch.cat((embed.cos(), embed.sin()), dim=-1)


INIT_ARGS = {
    'k': 3,
    'include_input': True,
    'use_logspace': True,
    'max_freq': None,
}

ARGS = {
    'x': torch.randn(2, 4),
}

if __name__ == '__main__':
    model = FourierFeatEnc(**INIT_ARGS)
    output = model(**ARGS)
    print(f"Input shape: {ARGS['x'].shape}")
    print(f"Output shape: {output.shape}")
    print(f"Output:\n{output}")
Setup block drafted by
a language model, class construct-then-call prompt, model not recorded
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 2, dtype family float, shape (2, 28) — shape (2, 28) 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: torch and numpy seeded to 0 by the harness immediately before the class was constructed from INIT_ARGS (fixture_check.py)
Mapping2Dto3D Ran
ran on a synthesized input
tum-vision/intrinsic-neural-fields/neutex/neutex.py
code served (permissive licence) · get_code("c4ab41810ab166aa")
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 Tensor, ndim 2, 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
import torch.nn as nn
import torch.nn.functional as F
import numpy as np
from torch.nn import init

def get_xavier_multiplier(m, gain):
    if isinstance(m, nn.Conv1d):
        ksize = m.kernel_size[0]
        n1 = m.in_channels
        n2 = m.out_channels
        std = gain * np.sqrt(2.0 / ((n1 + n2) * ksize))
    elif isinstance(m, nn.ConvTranspose1d):
        ksize = m.kernel_size[0] // m.stride[0]
        n1 = m.in_channels
        n2 = m.out_channels
        std = gain * np.sqrt(2.0 / ((n1 + n2) * ksize))
    elif isinstance(m, nn.Conv2d):
        ksize = m.kernel_size[0] * m.kernel_size[1]
        n1 = m.in_channels
        n2 = m.out_channels
        std = gain * np.sqrt(2.0 / ((n1 + n2) * ksize))
    elif isinstance(m, nn.ConvTranspose2d):
        ksize = m.kernel_size[0] * m.kernel_size[1] // m.stride[0] // m.stride[1]
        n1 = m.in_channels
        n2 = m.out_channels
        std = gain * np.sqrt(2.0 / ((n1 + n2) * ksize))
    elif isinstance(m, nn.Conv3d):
        ksize = m.kernel_size[0] * m.kernel_size[1] * m.kernel_size[2]
        n1 = m.in_channels
        n2 = m.out_channels
        std = gain * np.sqrt(2.0 / ((n1 + n2) * ksize))
    elif isinstance(m, nn.ConvTranspose3d):
        ksize = m.kernel_size[0] * m.kernel_size[1] * m.kernel_size[2] // m.stride[0] // m.stride[1] // m.stride[2]
        n1 = m.in_channels
        n2 = m.out_channels
        std = gain * np.sqrt(2.0 / ((n1 + n2) * ksize))
    elif isinstance(m, nn.Linear):
        n1 = m.in_features
        n2 = m.out_features
        std = gain * np.sqrt(2.0 / (n1 + n2))
    else:
        return None
    return std

def init_weights(net, init_type='xavier_uniform', gain=1):

    def init_func(m):
        classname = m.__class__.__name__
        if hasattr(m, 'weight') and (classname.find('Conv') != -1 or classname.find('Linear') != -1):
            if init_type == 'xavier_uniform':
                xavier_uniform_(m, gain)
            elif init_type == 'normal':
                init.normal_(m.weight.data, 0.0, gain)
            elif init_type == 'xavier':
                init.xavier_normal_(m.weight.data, gain=gain)
            elif init_type == 'kaiming':
                init.kaiming_normal_(m.weight.data, a=0, mode='fan_in')
            elif init_type == 'orthogonal':
                init.orthogonal_(m.weight.data, gain=gain)
            else:
                raise NotImplementedError('initialization method [{}] is not implemented'.format(init_type))
            if hasattr(m, 'bias') and m.bias is not None:
                init.constant_(m.bias.data, 0.0)
        elif classname.find('BatchNorm2d') != -1:
            init.normal_(m.weight.data, 1.0, gain)
            init.constant_(m.bias.data, 0.0)
    net.apply(init_func)

def xavier_uniform_(m, gain):
    std = get_xavier_multiplier(m, gain)
    m.weight.data.uniform_(-std * np.sqrt(3.0), std * np.sqrt(3.0))

class Mapping2Dto3D(nn.Module):
    """
    Modified AtlasNet core function
    """

    def __init__(self, input_point_dim, hidden_size=512, num_layers=2, activation='relu'):
        """
        template_size: input size
        """
        super().__init__()
        assert activation in ['relu', 'softplus']
        self.input_size = input_point_dim
        self.dim_output = 3
        self.hidden_neurons = hidden_size
        self.num_layers = num_layers
        self.linear1 = nn.Linear(self.input_size, self.hidden_neurons)
        init_weights(self.linear1)
        self.linear_list = nn.ModuleList([nn.Linear(self.hidden_neurons, self.hidden_neurons) for _ in range(self.num_layers)])
        for l in self.linear_list:
            init_weights(l)
        self.last_linear = nn.Linear(self.hidden_neurons, self.dim_output)
        init_weights(self.last_linear)
        if activation == 'relu':
            self.activation = F.relu
        elif activation == 'softplus':
            self.activation = F.softplus

    def _forward(self, x):
        x = self.activation(self.linear1(x))
        for i in range(self.num_layers):
            x = self.activation(self.linear_list[i](x))
        return self.last_linear(x)

    def forward(self, x):
        assert x.shape[-1] == self.input_size
        self.xyz_ = self._forward(x)
        return self.xyz_

INIT_ARGS = {
    'input_point_dim': 4,
    'hidden_size': 64,
    'num_layers': 2,
    'activation': 'relu'
}

ARGS = {
    'x': torch.randn(2, 4)
}

if __name__ == '__main__':
    model = Mapping2Dto3D(**INIT_ARGS)
Setup block drafted by
a language model, class construct-then-call prompt, model not recorded
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 2, dtype family float — no output shape is stated: observe_signature.py observed this class with its constructor sizes canonicalised where it could, which the node does not record, so its shape is not stated as this setup block's
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: torch and numpy seeded to 0 by the harness immediately before the class was constructed from INIT_ARGS (fixture_check.py)
MappingManifold Ran
ran on a synthesized input
tum-vision/intrinsic-neural-fields/neutex/neutex.py
code served (permissive licence) · get_code("2302d94a2911595e")
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 Tensor, ndim 2, 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
import torch.nn as nn
import torch.nn.functional as F
import numpy as np
from torch.nn import init

def get_xavier_multiplier(m, gain):
    if isinstance(m, nn.Conv1d):
        ksize = m.kernel_size[0]
        n1 = m.in_channels
        n2 = m.out_channels
        std = gain * np.sqrt(2.0 / ((n1 + n2) * ksize))
    elif isinstance(m, nn.ConvTranspose1d):
        ksize = m.kernel_size[0] // m.stride[0]
        n1 = m.in_channels
        n2 = m.out_channels
        std = gain * np.sqrt(2.0 / ((n1 + n2) * ksize))
    elif isinstance(m, nn.Conv2d):
        ksize = m.kernel_size[0] * m.kernel_size[1]
        n1 = m.in_channels
        n2 = m.out_channels
        std = gain * np.sqrt(2.0 / ((n1 + n2) * ksize))
    elif isinstance(m, nn.ConvTranspose2d):
        ksize = m.kernel_size[0] * m.kernel_size[1] // m.stride[0] // m.stride[1]
        n1 = m.in_channels
        n2 = m.out_channels
        std = gain * np.sqrt(2.0 / ((n1 + n2) * ksize))
    elif isinstance(m, nn.Conv3d):
        ksize = m.kernel_size[0] * m.kernel_size[1] * m.kernel_size[2]
        n1 = m.in_channels
        n2 = m.out_channels
        std = gain * np.sqrt(2.0 / ((n1 + n2) * ksize))
    elif isinstance(m, nn.ConvTranspose3d):
        ksize = m.kernel_size[0] * m.kernel_size[1] * m.kernel_size[2] // m.stride[0] // m.stride[1] // m.stride[2]
        n1 = m.in_channels
        n2 = m.out_channels
        std = gain * np.sqrt(2.0 / ((n1 + n2) * ksize))
    elif isinstance(m, nn.Linear):
        n1 = m.in_features
        n2 = m.out_features
        std = gain * np.sqrt(2.0 / (n1 + n2))
    else:
        return None
    return std

def init_weights(net, init_type='xavier_uniform', gain=1):

    def init_func(m):
        classname = m.__class__.__name__
        if hasattr(m, 'weight') and (classname.find('Conv') != -1 or classname.find('Linear') != -1):
            if init_type == 'xavier_uniform':
                xavier_uniform_(m, gain)
            elif init_type == 'normal':
                init.normal_(m.weight.data, 0.0, gain)
            elif init_type == 'xavier':
                init.xavier_normal_(m.weight.data, gain=gain)
            elif init_type == 'kaiming':
                init.kaiming_normal_(m.weight.data, a=0, mode='fan_in')
            elif init_type == 'orthogonal':
                init.orthogonal_(m.weight.data, gain=gain)
            else:
                raise NotImplementedError('initialization method [{}] is not implemented'.format(init_type))
            if hasattr(m, 'bias') and m.bias is not None:
                init.constant_(m.bias.data, 0.0)
        elif classname.find('BatchNorm2d') != -1:
            init.normal_(m.weight.data, 1.0, gain)
            init.constant_(m.bias.data, 0.0)
    net.apply(init_func)

def xavier_uniform_(m, gain):
    std = get_xavier_multiplier(m, gain)
    m.weight.data.uniform_(-std * np.sqrt(3.0), std * np.sqrt(3.0))

class MappingManifold(nn.Module):

    def __init__(self, input_dim, output_dim, hidden_size=128, num_layers=2):
        """
        template_size: input size
        """
        super().__init__()
        self.input_dim = input_dim
        self.output_dim = output_dim
        self.hidden_neurons = hidden_size
        self.num_layers = num_layers
        self.linear1 = nn.Linear(self.input_dim, self.hidden_neurons)
        init_weights(self.linear1)
        self.linear_list = nn.ModuleList([nn.Linear(self.hidden_neurons, self.hidden_neurons) for _ in range(self.num_layers)])
        for l in self.linear_list:
            init_weights(l)
        self.last_linear = nn.Linear(self.hidden_neurons, self.output_dim)
        init_weights(self.last_linear)
        self.activation = F.relu

    def forward(self, x):
        x = self.activation(self.linear1(x))
        for i in range(self.num_layers):
            x = self.activation(self.linear_list[i](x))
        return self.last_linear(x)


INIT_ARGS = {
    'input_dim': 4,
    'output_dim': 3,
    'hidden_size': 8,
    'num_layers': 2,
}

ARGS = {
    'x': torch.randn(2, 4),
}

if __name__ == '__main__':
    model = MappingManifold(**INIT_ARGS)
    output = model(**ARGS)
    print(f"Output shape: {output.shape}")
    print(f"Output: {output}")
Setup block drafted by
a language model, class construct-then-call prompt, model not recorded
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 2, dtype family float — no output shape is stated: observe_signature.py observed this class with its constructor sizes canonicalised where it could, which the node does not record, so its shape is not stated as this setup block's
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: torch and numpy seeded to 0 by the harness immediately before the class was constructed from INIT_ARGS (fixture_check.py)
SphereTemplate Ran
ran on a synthesized input
tum-vision/intrinsic-neural-fields/neutex/neutex.py
code served (permissive licence) · get_code("13ead961d06dcb59")
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 Tensor, ndim 2, dtype family float, shape (8, 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 torch.nn.functional as F

class SphereTemplate:

    def get_random_points(self, npoints, device):
        """
        Get random points on a Sphere
        Return Tensor of Size [x, 3, x ... x]
        """
        with torch.no_grad():
            points = torch.randn((npoints, 3)).to(device).float() * 2 - 1
            points = F.normalize(points, dim=-1)
        return points

    def get_regular_points(self, npoints, device):
        """
        Get regular points on a Sphere
        Return Tensor of Size [x, 3]
        """
        import trimesh
        mesh = trimesh.creation.icosphere(6)
        return torch.tensor(mesh.vertices).to(device).float()


INIT_ARGS = {}

ARGS = {
    'npoints': 8,
    'device': 'cpu'
}

METHOD = 'get_random_points'

if __name__ == '__main__':
    sphere = SphereTemplate(**INIT_ARGS)
    result = getattr(sphere, METHOD)(**ARGS)
    print(f"Result shape: {result.shape}")
    print(f"Result dtype: {result.dtype}")
    print(f"Result:\n{result}")
Setup block drafted by
a language model, class construct-then-call prompt, model not recorded
Imports
torch, trimesh — import names as written, not package names to install; a name the repository itself provides is not a package
Returned
type Tensor, ndim 2, dtype family float, shape (8, 3) — shape (8, 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: torch and numpy seeded to 0 by the harness immediately before the class was constructed from INIT_ARGS (fixture_check.py)
SquareTemplate Ran
ran on a synthesized input
tum-vision/intrinsic-neural-fields/neutex/neutex.py
code served (permissive licence) · get_code("463c1518a80f4402")
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 Tensor, ndim 2, dtype family float, shape (16, 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

class SquareTemplate:

    def __init__(self):
        self.regular_num_points = 0

    def get_random_points(self, npoints, device):
        with torch.no_grad():
            rand_grid = (torch.rand((npoints, 2)) * 2 - 1).to(device).float()
            return rand_grid

    def get_regular_points(self, npoints=2500, device=None):
        """
        Get regular points on a Square
        """
        assert int(npoints ** 0.5) ** 2 == npoints
        assert device is not None, 'device needs to be provided for get_regular_points'
        side_length = int(npoints ** 0.5)
        uv = np.stack(np.meshgrid(*[np.linspace(-1, 1, side_length)] * 2, indexing='ij'), axis=-1).reshape((-1, 2))
        points = torch.FloatTensor(uv).to(device)
        return points.requires_grad_()


INIT_ARGS = {}

ARGS = {
    'npoints': 16,
    'device': 'cpu'
}

METHOD = 'get_random_points'

if __name__ == '__main__':
    template = SquareTemplate(**INIT_ARGS)
    result = getattr(template, METHOD)(**ARGS)
    print("Random points shape:", result.shape)
    print("Random points sample:\n", result[:3])
    
    regular_result = template.get_regular_points(npoints=16, device='cpu')
    print("Regular points shape:", regular_result.shape)
    print("Regular points sample:\n", regular_result[:3])
Setup block drafted by
a language model, class construct-then-call prompt, model not recorded
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 2, dtype family float, shape (16, 2) — shape (16, 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: torch and numpy seeded to 0 by the harness immediately before the class was constructed from INIT_ARGS (fixture_check.py)
get_xavier_multiplier Ran
ran, honoured the contract we drafted
tum-vision/intrinsic-neural-fields/neutex/neutex.py
code served (permissive licence) · get_code("76fd3b77e741b50d")
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 float64, ndim 0, 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
import torch.nn as nn
import numpy as np

torch.manual_seed(0)

# Test with Conv1d
m_conv1d = nn.Conv1d(in_channels=3, out_channels=5, kernel_size=7)
gain_value = 1.0

ARGS = {
    "m": m_conv1d,
    "gain": gain_value
}
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 float64, ndim 0, dtype family float — no output shape is stated: observe_signature.py recorded float ndim 0, which disagrees with the fixture run
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
Atlasnet Not yet run
unverified (harvested, no recorded run)
tum-vision/intrinsic-neural-fields/neutex/neutex.py
code served (permissive licence) · get_code("4741580d189cfcf3")
How it ran

no invocation recorded.

InverseAtlasnet Not yet run
unverified (harvested, no recorded run)
tum-vision/intrinsic-neural-fields/neutex/neutex.py
code served (permissive licence) · get_code("1463d061ca5bebd0")
How it ran

no invocation recorded.

NeuTex Not yet run
unverified (harvested, no recorded run)
tum-vision/intrinsic-neural-fields/neutex/neutex.py
code served (permissive licence) · get_code("26ae50a466262cc4")
How it ran

no invocation recorded.

TextureMlp Not yet run
unverified (harvested, no recorded run)
tum-vision/intrinsic-neural-fields/neutex/neutex.py
code served (permissive licence) · get_code("0deb2251d87892c9")
How it ran

no invocation recorded.

TextureMlpMix Not yet run
unverified (harvested, no recorded run)
tum-vision/intrinsic-neural-fields/neutex/neutex.py
code served (permissive licence) · get_code("3de5f2fdd66c53b6")
How it ran

no invocation recorded.

init_seq Not yet run
unverified (harvested, no recorded run)
tum-vision/intrinsic-neural-fields/neutex/neutex.py
code served (permissive licence) · get_code("bfb53d55b842d271")
How it ran

no invocation recorded.

init_weights Not yet run
unverified (harvested, no recorded run)
tum-vision/intrinsic-neural-fields/neutex/neutex.py
code served (permissive licence) · get_code("865220c86d035cad")
How it ran

no invocation recorded.

xavier_uniform_ Not yet run
unverified (harvested, no recorded run)
tum-vision/intrinsic-neural-fields/neutex/neutex.py
code served (permissive licence) · get_code("f915ebd2b117d242")
How it ran

no invocation recorded.

Also present in other papers' code

Of 14 samples lifted from this paper's linked repositories, 14 carry a function identity and 0 do not. Of the 14 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 14 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

Neural fields have gained significant attention in the computer vision community due to their excellent performance in novel view synthesis, geometry reconstruction, and generative modeling. Some of their advantages are a sound theoretic foundation and an easy implementation in current deep learning frameworks. While neural fields have been applied to signals on manifolds, e.g., for texture reconstruction, their representation has been limited to extrinsically embedding the shape into Euclidean space. The extrinsic embedding ignores known intrinsic manifold properties and is inflexible wrt. transfer of the learned function. To overcome these limitations, this work introduces intrinsic neural fields, a novel and versatile representation for neural fields on manifolds. Intrinsic neural fields combine the advantages of neural fields with the spectral properties of the Laplace-Beltrami operator. We show theoretically that intrinsic neural fields inherit many desirable properties of the extrinsic neural field framework but exhibit additional intrinsic qualities, like isometry invariance. In experiments, we show intrinsic neural fields can reconstruct high-fidelity textures from images with state-of-the-art quality and are robust to the discretization of the underlying manifold. We demonstrate the versatility of intrinsic neural fields by tackling various applications: texture transfer between deformed shapes & different shapes, texture reconstruction from real-world images with view dependence, and discretization-agnostic learning on meshes and point clouds.

For agents

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

get_harvested_code_for_paper("2203.07967")
get_code_for_paper("2203.07967")
have("2203.07967")

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

curl https://syntology.ai/api/ran/2203.07967.json

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

[![Syntology run record](https://syntology.ai/api/ran/2203.07967.svg)](https://syntology.ai/paper/2203.07967)

Connect an agent — have() is free.