SYNTOLOGY HomeExplorerAtlasCodeMethodologyAboutDevelopersFeedPricing
Paper · 2410.09299 · ICLR · 2025

Hierarchical uncertainty estimation for learning-based registration in neuroimaging

Peirong Liu, Xiaoling Hu, Karthik Gopinath, Malte Hoffmann, Koen Leemput, Oula Puonti, Juan Iglesias

arXiv · PDF · Open in the Atlas

Code that ran

13 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; 2 honoured the contract we drafted, 1 violated it and 0 have no contract verdict; 10 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/2410.09299.json · badge: /api/ran/2410.09299.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
HuXiaoling/Regre4Regis canonical (classifier label)
classifier basis: census
11 ran · 1 no recorded run
huxiaoling/regre4regis canonical (classifier label)
classifier basis: classifier_2025
2 ran · 0 no recorded run
FunctionStatusWhere it lives
conv_block_2_3d Ran
ran; our contract draft was wrong, not the code
HuXiaoling/Regre4Regis/unet/unet_3d.py
code served (permissive licence) · get_code("769628dc3a9af468")
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)

# Define activation function
activation = nn.ReLU()

# Define parameters for conv_block_2_3d
in_dim = 3
out_dim = 5

ARGS = {
    'in_dim': in_dim,
    'out_dim': out_dim,
    'activation': activation
}
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
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_block_3d Ran
ran; our contract draft was wrong, not the code
HuXiaoling/Regre4Regis/unet/unet_3d.py
code served (permissive licence) · get_code("520950cb24a9041a")
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 = {
    'in_dim': 3,
    'out_dim': 16,
    'activation': nn.ReLU()
}
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_trans_block_3d Ran
ran; our contract draft was wrong, not the code
HuXiaoling/Regre4Regis/unet/unet_3d.py
code served (permissive licence) · get_code("74e9ef01fdad6274")
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)

# Define arguments for conv_trans_block_3d
in_dim = 16
out_dim = 32
activation = nn.ReLU()

ARGS = {
    "in_dim": in_dim,
    "out_dim": out_dim,
    "activation": activation
}
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
dice_loss Ran
ran, violated the contract we drafted
HuXiaoling/Regre4Regis/utilities.py
code served (permissive licence) · get_code("398c327fcb4cf242")
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. 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)

# Create small binary segmentation predictions and targets
# The function reshapes to 1D and computes dice loss
input_tensor = torch.tensor([
    [[0.9, 0.1, 0.8, 0.2],
     [0.7, 0.3, 0.6, 0.4]],
    [[0.85, 0.15, 0.75, 0.25],
     [0.65, 0.35, 0.55, 0.45]]
], dtype=torch.float32)

target_tensor = torch.tensor([
    [[1, 0, 1, 0],
     [1, 0, 1, 0]],
    [[1, 0, 1, 0],
     [1, 0, 1, 0]]
], dtype=torch.float32)

ARGS = {
    'input': input_tensor,
    'target': target_tensor
}
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
fast_3D_interp_torch Ran
ran; the contract check could not run (RAISES, DEP_MISSING or TIMEOUT)
HuXiaoling/Regre4Regis/fitting_functions.py
code served (permissive licence) · get_code("3640127790f3d5d1")
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 2, dtype family float, shape (2, 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)

# Create a small 3D volume with shape (4, 3, 5) - distinct sizes for each dimension
# Add a channel dimension to make it (4, 3, 5, 1)
X = torch.randn(4, 3, 5, 1)

# Create coordinate tensors for interpolation
# Shape (2, 2) for batch of 2x2 query points
II = torch.tensor([[0.5, 1.5], [2.5, 3.0]], dtype=torch.float32)
JJ = torch.tensor([[0.5, 1.0], [1.5, 2.0]], dtype=torch.float32)
KK = torch.tensor([[1.0, 2.0], [3.0, 4.0]], dtype=torch.float32)

mode = 'linear'
device = 'cpu'
default_value_linear = 0.0

ARGS = {
    'X': X,
    'II': II,
    'JJ': JJ,
    'KK': KK,
    'mode': mode,
    'device': device,
    'default_value_linear': default_value_linear
}
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 2, dtype family float, shape (2, 2) — shape (2, 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
get_tp_fp_fn_tn Ran
ran; our contract draft was wrong, not the code
HuXiaoling/Regre4Regis/utilities.py
code served (permissive licence) · get_code("89771afaf563e5e6")
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 3, dtype family float; no output shape is stated. 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 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)

# Create small tensors for testing
# net_output: (b, c, x, y) - network output predictions
batch_size = 2
num_classes = 3
spatial_x = 4
spatial_y = 5

net_output = torch.rand(batch_size, num_classes, spatial_x, spatial_y)

# gt: ground truth - can be label map (b, x, y) or one-hot (b, c, x, y)
# Using label map format (b, x, y)
gt = torch.randint(0, num_classes, (batch_size, spatial_x, spatial_y))

# axes: tuple of axes to sum over (spatial dimensions)
axes = (2, 3)

# mask: optional mask with shape (b, 1, x, y)
mask = torch.ones(batch_size, 1, spatial_x, spatial_y)

# square: whether to square the tp, fp, fn, tn values
square = False

ARGS = {
    "net_output": net_output,
    "gt": gt,
    "axes": axes,
    "mask": mask,
    "square": square,
}
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 tuple, ndim 3, 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
jaccard Ran
ran, honoured the contract we drafted
HuXiaoling/Regre4Regis/metrics.py
code served (permissive licence) · get_code("12b5ba241c136bb3")
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 float, 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 numpy as np

# Create test and reference numpy arrays with matching shapes
test = np.array([
    [[1, 0, 1, 0],
     [0, 1, 0, 1]],
    [[1, 1, 0, 0],
     [0, 0, 1, 1]]
], dtype=np.int32)

reference = np.array([
    [[1, 0, 0, 0],
     [0, 1, 0, 0]],
    [[1, 0, 0, 0],
     [0, 0, 1, 0]]
], dtype=np.int32)

# Create a ConfusionMatrix instance
class ConfusionMatrix:
    def __init__(self, test=None, reference=None):
        self.tp = None
        self.fp = None
        self.tn = None
        self.fn = None
        self.size = None
        self.reference_empty = None
        self.reference_full = None
        self.test_empty = None
        self.test_full = None
        self.set_reference(reference)
        self.set_test(test)

    def set_test(self, test):
        self.test = test
        self.reset()

    def set_reference(self, reference):
        self.reference = reference
        self.reset()

    def reset(self):
        self.tp = None
        self.fp = None
        self.tn = None
        self.fn = None
        self.size = None
        self.test_empty = None
        self.test_full = None
        self.reference_empty = None
        self.reference_full = None

    def compute(self):
        if self.test is None or self.reference is None:
            raise ValueError("'test' and 'reference' must both be set to compute confusion matrix.")
        self.tp = int(((self.test != 0) * (self.reference != 0)).sum())
        self.fp = int(((self.test != 0) * (self.reference == 0)).sum())
        self.tn = int(((self.test == 0) * (self.reference == 0)).sum())
        self.fn = int(((self.test == 0) * (self.reference != 0)).sum())
        self.size = int(np.prod(self.reference.shape, dtype=np.int64))
        self.test_empty = not np.any(self.test)
        self.test_full = np.all(self.test)
        self.reference_empty = not np.any(self.reference)
        self.reference_full = np.all(self.reference)

    def get_matrix(self):
        for entry in (self.tp, self.fp, self.tn, self.fn):
            if entry is None:
                self.compute()
                break
        return (self.tp, self.fp, self.tn, self.fn)

    def get_size(self):
        if self.size is None:
            self.compute()
        return self.size

    def get_existence(self):
        for case in (self.test_empty, self.test_full, self.reference_empty, self.reference_full):
            if case is None:
                self.compute()
                break
        return (self.test_empty, self.test_full, self.reference_empty, self.reference_full)

confusion_matrix = ConfusionMatrix(test, reference)

ARGS = {
    'test': None,
    'reference': None,
    'confusion_matrix': confusion_matrix,
    'nan_for_nonexisting': True
}
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 float, 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
make_gaussian_kernel Ran
ran, honoured the contract we drafted
HuXiaoling/Regre4Regis/fitting_functions.py
code served (permissive licence) · get_code("db1ce499dc61eda0")
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 float, shape (17,) 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 make_gaussian_kernel
ARGS = {
    "sigma": 2.5,
    "device": "cpu"
}
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 float, shape (17,) — shape (17,) 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
parse_func Ran
ran; our contract draft was wrong, not the code
huxiaoling/regre4regis/main_end2end.py
code served (permissive licence) · get_code("1d7a2af93fda68e6")
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 argparse
import json
import tempfile
import os

# Create a temporary params JSON file
params_dict = {
    "common": {
        "activity": "train",
        "dataset": "test_dataset",
        "num_classes": "10",
        "img_file": "images.txt",
        "checkpoint_restore": "checkpoint.pth"
    },
    "train": {
        "train_datalist": "train_list.txt",
        "validation_datalist": "val_list.txt",
        "mode": "segmentation",
        "regress_loss": "mse",
        "uncer": "True",
        "nonlin": "relu",
        "fitting": "standard",
        "output_folder": "/tmp/output",
        "loss_weight_mask": "1.0",
        "loss_weight_uncer": "0.5",
        "loss_weight_seg": "1.0",
        "num_workers": "4",
        "train_batch_size": "32",
        "validation_batch_size": "16",
        "learning_rate": "0.001",
        "num_epochs": "100",
        "save_every": "10"
    }
}

# Write to temporary file
temp_file = tempfile.NamedTemporaryFile(mode='w', suffix='.json', delete=False)
json.dump(params_dict, temp_file)
temp_file.close()

# Create argparse Namespace object
args = argparse.Namespace(params=temp_file.name)

ARGS = {
    'args': args
}
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
parse_func Ran
ran; our contract draft was wrong, not the code
huxiaoling/regre4regis/inference.py
code served (permissive licence) · get_code("ff7f71df6fdb4f7b")
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 argparse
import json
import tempfile
import os

# Create a temporary directory and params file
temp_dir = tempfile.mkdtemp()
params_file = os.path.join(temp_dir, 'params.json')

# Create the params JSON structure that the function expects
params = {
    'common': {
        'activity': 'validation',
        'dataset': 'cifar10',
        'num_classes': '10',
        'img_file': ['image1.jpg', 'image2.jpg'],
        'checkpoint_restore': '/path/to/checkpoint.pth'
    },
    'validation': {
        'validation_datalist': ['val_image1.jpg', 'val_image2.jpg'],
        'output_folder': '/path/to/output',
        'batch_size': '32'
    }
}

# Write the params file
with open(params_file, 'w') as f:
    json.dump(params, f)

# Create the args object
args = argparse.Namespace(params=params_file)

ARGS = {'args': args}
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
parse_func Ran
ran; our contract draft was wrong, not the code
HuXiaoling/Regre4Regis/main_uncer.py
code served (permissive licence) · get_code("132aeec83f85c38c")
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 json
import tempfile
import os

# Create a temporary params JSON file
params_dict = {
    "common": {
        "activity": "train",
        "dataset": "test_dataset",
        "num_classes": "10",
        "img_file": ["img1.jpg", "img2.jpg"],
        "checkpoint_restore": "checkpoint.pth"
    },
    "train": {
        "train_datalist": "train_list.txt",
        "validation_datalist": "val_list.txt",
        "mode": "segmentation",
        "regress_loss": "mse",
        "uncer": "bayesian",
        "output_folder": "./output",
        "loss_weight_mask": "1.0",
        "loss_weight_uncer": "0.5",
        "num_workers": "4",
        "train_batch_size": "32",
        "validation_batch_size": "16",
        "learning_rate": "0.001",
        "num_epochs": "50",
        "save_every": "10"
    }
}

# Write to a temporary file
temp_file = tempfile.NamedTemporaryFile(mode='w', suffix='.json', delete=False)
json.dump(params_dict, temp_file)
temp_file.close()

# Create argparse Namespace object with the params file path
import argparse
args = argparse.Namespace(params=temp_file.name)

ARGS = {
    'args': args
}
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 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
precision Ran
ran; our contract draft was wrong, not the code
HuXiaoling/Regre4Regis/metrics.py
code served (permissive licence) · get_code("035dc40fdafe0275")
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 float, 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 numpy as np

# Create test and reference numpy arrays
# They need to be the same shape for the assertion to pass
test_array = np.array([[1, 0, 1, 0],
                       [1, 1, 0, 0]])

reference_array = np.array([[1, 0, 1, 1],
                            [1, 0, 0, 0]])

# Create a ConfusionMatrix instance
confusion_matrix_obj = None  # Will be None initially, letting the function create it

ARGS = {
    'test': test_array,
    'reference': reference_array,
    'confusion_matrix': confusion_matrix_obj,
    'nan_for_nonexisting': True
}
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 float, 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
sum_tensor Ran
ran; the contract check could not run (RAISES, DEP_MISSING or TIMEOUT)
HuXiaoling/Regre4Regis/utilities.py
pointer only (licence: MIT); the stored copy get_code serves, from szc19990412/TransMIL, is pointer only (licence: NONE) · get_code("9e4e02bee8e90bea")
How it ran

The body of this sample is withheld: this paper's link allows the body inline (MIT), but get_code serves the stored copy, first harvested from szc19990412/TransMIL, 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.

dice Not yet run
unverified (harvested, no recorded run)
HuXiaoling/Regre4Regis/metrics.py
code served (permissive licence) · get_code("a85d8587c247d950")
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, 8 are also present in other papers whose code we hold (7 with at least one paper dated earlier, 0 the same month, 4 later) and 6 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
dice 4 other papers
2 dated earlier, 2 later
1809.10486 (2018-09)
nnU-Net: Self-adapting Framework for U-Net-Based Medical Image Segmentation
2212.04497 (2022-12)
UNETR++: Delving into Efficient and Accurate 3D Medical Image Segmentation
2506.23468 (2025-06)
NavM rph: A Self-Evolving World Model for Vision-and-Language Navigation in Continuous Environments
and 1 more not named here
jaccard 4 other papers
2 dated earlier, 2 later
1809.10486 (2018-09)
nnU-Net: Self-adapting Framework for U-Net-Based Medical Image Segmentation
2212.04497 (2022-12)
UNETR++: Delving into Efficient and Accurate 3D Medical Image Segmentation
2506.23468 (2025-06)
NavM rph: A Self-Evolving World Model for Vision-and-Language Navigation in Continuous Environments
and 1 more not named here
precision 4 other papers
2 dated earlier, 2 later
1809.10486 (2018-09)
nnU-Net: Self-adapting Framework for U-Net-Based Medical Image Segmentation
2212.04497 (2022-12)
UNETR++: Delving into Efficient and Accurate 3D Medical Image Segmentation
2506.23468 (2025-06)
NavM rph: A Self-Evolving World Model for Vision-and-Language Navigation in Continuous Environments
and 1 more not named here
sum_tensor 2 other papers
1 later, 1 same year, order not determinable
aaai_28394 (2024 (year only))
DTMFormer: Dynamic Token Merging for Boosting Transformer-Based Medical Image Segmentation
2602.01951 (2026-02)
Enabling Progressive Whole-slide Image Analysis with Multi-scale Pyramidal Network
conv_block_2_3d 1 other paper
1 dated earlier
1606.06650 (2016-06)
3D U-Net: Learning Dense Volumetric Segmentation from Sparse Annotation
conv_block_3d 1 other paper
1 dated earlier
1606.06650 (2016-06)
3D U-Net: Learning Dense Volumetric Segmentation from Sparse Annotation
conv_trans_block_3d 1 other paper
1 dated earlier
1606.06650 (2016-06)
3D U-Net: Learning Dense Volumetric Segmentation from Sparse Annotation
dice_loss 1 other paper
1 dated earlier
Yue_Connecting_the_Dots_Floorplan_Reconstruction_Using_Two-Level_Queries_CVPR_2023_paper (2023 (year only))
Connecting the Dots: Floorplan Reconstruction Using Two-Level Queries

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

Over recent years, deep learning based image registration has achieved impressive accuracy in many domains, including medical imaging and, specifically, human neuroimaging with magnetic resonance imaging (MRI). However, the uncertainty estimation associated with these methods has been largely limited to the application of generic techniques (e.g., Monte Carlo dropout) that do not exploit the peculiarities of the problem domain, particularly spatial modeling. Here, we propose a principled way to propagate uncertainties (epistemic or aleatoric) estimated at the level of spatial location by these methods, to the level of global transformation models, and further to downstream tasks. Specifically, we justify the choice of a Gaussian distribution for the local uncertainty modeling, and then propose a framework where uncertainties spread across hierarchical levels, depending on the choice of transformation model. Experiments on publicly available data sets show that Monte Carlo dropout correlates very poorly with the reference registration error, whereas our uncertainty estimates correlate much better. Crucially, the results also show that uncertainty-aware fitting of transformations improves the registration accuracy of brain MRI scans. Finally, we illustrate how sampling from the posterior distribution of the transformations can be used to propagate uncertainties to downstream neuroimaging tasks. Code is available at: https://github.com/HuXiaoling/Regre4Regis.

For agents

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

get_harvested_code_for_paper("2410.09299")
get_code_for_paper("2410.09299")
have("2410.09299")

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

curl https://syntology.ai/api/ran/2410.09299.json

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

[![Syntology run record](https://syntology.ai/api/ran/2410.09299.svg)](https://syntology.ai/paper/2410.09299)

Connect an agent — have() is free.