SYNTOLOGY HomeExplorerAtlasCodeMethodologyAboutDevelopersFeedPricing
Paper · 2602.03555 · 2026

Cut to the Mix: Simple Data Augmentation Outperforms Elaborate Ones in Limited Organ Segmentation Datasets

Chang Liu, Andreas Maier, Fuxin Fan, Annette Schwarz

arXiv · PDF · Open in the Atlas

Code that ran

6 of 7 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, 1 violated it and 0 have no contract verdict; 5 failures of our instrument; 1 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/2602.03555.json · badge: /api/ran/2602.03555.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
Rebooorn/mosDAtoolkit canonical (classifier label)
classifier basis: census
6 ran · 1 no recorded run
FunctionStatusWhere it lives
augment_plan Ran
ran; our contract draft was wrong, not the code
Rebooorn/mosDAtoolkit/cli_carvemix.py
code served (permissive licence) · get_code("0f9bb8b23c90a674")
How it ran

Ran in Syntology's sandbox on this synthesized input (drafted by a language model, Claude Haiku 4.5, prompt python_fixture_v1, through the batch API) and returned list, ndim 1, dtype family float; no output shape is stated. The image it ran in was not recorded. If your copy returns the same type and ndim on these inputs, it matched what we observed; nothing here checks that. It does NOT mean the output is correct: nothing has compared it to the paper.

import random

# Analyze the function:
# - train_list: used with rd.choice(fnames), so needs to be a list of items
# - nclass: used in range(nclass), so needs to be an int
# - num_augment: used in range(num_augment), so needs to be an int (default 500)

# train_list should be a list of filenames/strings
train_list = ['image_001.jpg', 'image_002.jpg', 'image_003.jpg', 'image_004.jpg', 'image_005.jpg']

nclass = 3

num_augment = 10

ARGS = {
    'train_list': train_list,
    'nclass': nclass,
    'num_augment': num_augment
}
Setup block drafted by
a language model, Claude Haiku 4.5, prompt python_fixture_v1, through the batch API
Imports
none outside the standard library — import names as written, not package names to install; a name the repository itself provides is not a package
Returned
type list, ndim 1, dtype family float — a list output has no array shape of its own; none is stated
dtype family
int means an integer dtype; float means any other dtype the instrument saw, bool and complex included (fixture_check.py records only these two families)
Environment
image not recorded; re-verify against syntology-harvest-sandbox:v5-deps; not recorded for this sample: the harness installs a cuda-to-cpu shim (cuda_shim.py) before every sample and no run records whether a sample needed it; seed: none by the harness; any seed is in the setup block
extract_holes Ran
ran, violated the contract we drafted
Rebooorn/mosDAtoolkit/cli_anatomix.py
code served (permissive licence) · get_code("6fd6b3e1612690d6")
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 float, shape (2, 3, 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

# Create sample masks for the extract_holes function
# orig_mask: original mask with some positive values (object regions)
# aug_mask: augmented mask that may have zeros (background) where orig_mask had objects

orig_mask = np.array([
    [[0, 1, 2, 0],
     [1, 2, 1, 0],
     [0, 1, 0, 0]],
    [[0, 0, 3, 1],
     [2, 1, 0, 0],
     [1, 0, 2, 1]]
], dtype=np.float32)

aug_mask = np.array([
    [[0, 1, 0, 0],
     [1, 2, 0, 0],
     [0, 0, 0, 0]],
    [[0, 0, 3, 0],
     [0, 1, 0, 0],
     [1, 0, 0, 1]]
], dtype=np.float32)

ARGS = {
    "orig_mask": orig_mask,
    "aug_mask": aug_mask
}
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 3, dtype family float, shape (2, 3, 4) — shape (2, 3, 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
get_distance Ran
ran; the contract check could not run (RAISES, DEP_MISSING or TIMEOUT)
Rebooorn/mosDAtoolkit/cli_carvemix.py
code served (permissive licence) · get_code("373014fa9c9c5f9c")
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 (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 numpy as np
from scipy import ndimage

# Create a simple binary mask (foreground/background)
# Using a small 4x3 array for distinctness
f = np.array([
    [1, 0, 1],
    [0, 1, 0],
    [1, 1, 0],
    [0, 0, 1]
], dtype=bool)

# Create spacing for the distance transform
# Using a tuple of distinct values for each dimension
spacing = (1.0, 2.0)

ARGS = {
    "f": f,
    "spacing": spacing
}
Setup block drafted by
a language model, Claude Haiku 4.5, prompt python_fixture_v1, through the batch API
Imports
numpy, scipy — 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 (4, 3) — shape (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
pad_x_to_y_3d Ran
ran; our contract draft was wrong, not the code
Rebooorn/mosDAtoolkit/inpaint/net.py
code served (permissive licence) · get_code("57b331509573db94")
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, 3, 6, 8, 10) 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 two 5D tensors for 3D padding
# x will be smaller, y will be larger in dimensions 2, 3, 4
# Shape: (batch, channels, depth, height, width)
x = torch.randn(2, 3, 4, 5, 6)  # smaller tensor
y = torch.randn(2, 3, 6, 8, 10)  # larger tensor to pad x to match

ARGS = {
    "x": x,
    "y": y,
}
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, 3, 6, 8, 10) — shape (2, 3, 6, 8, 10) 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
total_variation_loss Ran
ran; the contract check could not run (RAISES, DEP_MISSING or TIMEOUT)
Rebooorn/mosDAtoolkit/inpaint/loss.py
code served (permissive licence) · get_code("dedfa1b69056bdc7")
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)

# The function expects a 4D tensor (batch, channels, height, width)
# It computes total variation loss by comparing adjacent pixels
# Using small but distinct dimensions: batch=2, channels=3, height=4, width=5
image = torch.randn(2, 3, 4, 5)

ARGS = {
    "image": image
}
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
weights_init Ran
ran; our contract draft was wrong, not the code
Rebooorn/mosDAtoolkit/inpaint/net.py
code served (permissive licence) · get_code("b0a0259ffdb9c7d8")
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 function. 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
import math

# Set seed for reproducibility
torch.manual_seed(0)

# Create a simple model with Conv and Linear layers to test the initialization
class SimpleModel(nn.Module):
    def __init__(self):
        super(SimpleModel, self).__init__()
        self.conv1 = nn.Conv2d(3, 16, kernel_size=3)
        self.linear1 = nn.Linear(16, 8)
        self.linear2 = nn.Linear(8, 4)
    
    def forward(self, x):
        x = self.conv1(x)
        x = x.view(x.size(0), -1)
        x = self.linear1(x)
        x = self.linear2(x)
        return x

# Define ARGS dict with all parameters for weights_init
ARGS = {
    'init_type': 'gaussian'
}
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 function — a function output has no array shape of its own; none is stated — the instrument coerces any output to an array; for a function 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
augment_plan Not yet run
unverified (harvested, no recorded run)
Rebooorn/mosDAtoolkit/cli_objectaug.py
code served (permissive licence) · get_code("a5654b577b228dce")
How it ran

no invocation recorded.

Also present in other papers' code

Of 7 samples lifted from this paper's linked repositories, 7 carry a function identity and 0 do not. Of the 7 distinct functions with an identity, 2 are also present in other papers whose code we hold (2 with at least one paper dated earlier, 0 the same month, 0 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
weights_init 8 other papers
8 dated earlier
1611.07004 (2016-11)
Image-to-Image Translation with Conditional Adversarial Networks
1712.02478 (2017-12)
Stacked Conditional Generative Adversarial Networks for Jointly Learning Shadow Detection and Shadow Removal
1804.07723 (2018-04)
Image Inpainting for Irregular Holes Using Partial Convolutions
and 5 more not named here
total_variation_loss 5 other papers
5 dated earlier
1611.07004 (2016-11)
Image-to-Image Translation with Conditional Adversarial Networks
1804.07723 (2018-04)
Image Inpainting for Irregular Holes Using Partial Convolutions
2004.02788 (2020-04)
Self-Supervised Scene De-occlusion
and 2 more not named here

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

Multi-organ segmentation is a widely applied clinical routine and automated organ segmentation tools dramatically improve the pipeline of the radiologists. Recently, deep learning (DL) based segmentation models have shown the capacity to accomplish such a task. However, the training of the segmentation networks requires large amount of data with manual annotations, which is a major concern due to the data scarcity from clinic. Working with limited data is still common for researches on novel imaging modalities. To enhance the effectiveness of DL models trained with limited data, data augmentation (DA) is a crucial regularization technique. Traditional DA (TDA) strategies focus on basic intra-image operations, i.e. generating images with different orientations and intensity distributions. In contrast, the interimage and object-level DA operations are able to create new images from separate individuals. However, such DA strategies are not well explored on the task of multi-organ segmentation. In this paper, we investigated four possible inter-image DA strategies: CutMix, CarveMix, ObjectAug and AnatoMix, on two organ segmentation datasets. The result shows that CutMix, CarveMix and AnatoMix can improve the average dice score by 4.9, 2.0 and 1.9, compared with the state-of-the-art nnUNet without DA strategies. These results can be further improved by adding TDA strategies. It is revealed in our experiments that Cut-Mix is a robust but simple DA strategy to drive up the segmentation performance for multi-organ segmentation, even when CutMix produces intuitively 'wrong' images. Our implementation is publicly available at https://github.com/Rebooorn/mosDAtoolkit for future benchmarks.

For agents

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

get_harvested_code_for_paper("2602.03555")
get_code_for_paper("2602.03555")
have("2602.03555")

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

curl https://syntology.ai/api/ran/2602.03555.json

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

[![Syntology run record](https://syntology.ai/api/ran/2602.03555.svg)](https://syntology.ai/paper/2602.03555)

Connect an agent — have() is free.