SYNTOLOGY HomeExplorerAtlasCodeMethodologyAboutDevelopersFeedPricing
Paper · 2601.21159 · 2026

Spatial-Regularization-Aware Dual-Branch Collaborative Inference for Training-Free OVSS in Remote Sensing Imagery

Jianzheng Wang, Huan Ni

arXiv · PDF · Open in the Atlas

Code that ran

11 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; 1 honoured the contract we drafted, 1 violated it and 0 have no contract verdict; 9 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/2601.21159.json · badge: /api/ran/2601.21159.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
yu-ni1989/SDCI canonical (classifier label)
classifier basis: census
11 ran · 1 no recorded run
FunctionStatusWhere it lives
cam_to_label Ran
ran; the contract check could not run (RAISES, DEP_MISSING or TIMEOUT)
yu-ni1989/SDCI/util/miou.py
code served (permissive licence) · get_code("28fbb08ac4592534")
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. 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

torch.manual_seed(0)

# Construct tiny tensors and arguments
# cam: shape (b, c, h, w) - batch=2, channels=4, height=3, width=5
cam = torch.randn(2, 4, 3, 5)

# cls_label: shape (b, c) - batch=2, channels=4
# Values should be floats that can be compared with cls_thre=0.4
cls_label = torch.tensor([[0.5, 0.3, 0.8, 0.2],
                          [0.6, 0.1, 0.9, 0.4]], dtype=torch.float32)

ARGS = {
    'cam': cam,
    'cls_label': cls_label,
    'bkg_thre': 0.3,
    'cls_thre': 0.4,
    'is_normalize': True
}
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 — a tuple 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
compute_ap Ran
ran; our contract draft was wrong, not the code
yu-ni1989/SDCI/util/tools.py
code served (permissive licence) · get_code("8b0474ce348169e2")
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 (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 torch

torch.manual_seed(0)

# predictions: shape (batch_size, num_class) - predicted scores for each class
# labels: shape (batch_size, num_class) - binary labels for each class
batch_size = 2
num_class = 4

predictions = torch.randn(batch_size, num_class)
labels = torch.randint(0, 2, (batch_size, num_class)).float()

ARGS = {
    "predictions": predictions,
    "labels": labels,
}
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 1, dtype family float, shape (4,) — shape (4,) as a later run recorded it
dtype family
int means an integer dtype; float means any other dtype the instrument saw, bool and complex included (fixture_check.py records only these two families)
Environment
image not recorded; re-verify against syntology-harvest-sandbox:v5-deps; not recorded for this sample: the harness installs a cuda-to-cpu shim (cuda_shim.py) before every sample and no run records whether a sample needed it; seed: none by the harness; any seed is in the setup block
compute_f1 Ran
ran; the contract check could not run (RAISES, DEP_MISSING or TIMEOUT)
yu-ni1989/SDCI/util/tools.py
code served (permissive licence) · get_code("c3562972f2ab9e25")
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

torch.manual_seed(0)

# Create small tensors for testing
batch_size = 2
num_classes = 4

# predictions: float tensor of shape (batch_size, num_classes)
# Values should be in a reasonable range for thresholding
predictions = torch.rand(batch_size, num_classes, dtype=torch.float32)

# labels: binary tensor of shape (batch_size, num_classes)
labels = torch.randint(0, 2, (batch_size, num_classes), dtype=torch.float32)

# mode_f1: one of 'overall', 'category', 'sample'
mode_f1 = 'overall'

# k_val: can be >= 1 (for topk) or < 1 (for relative/absolute threshold)
k_val = 2

# use_relative: boolean flag
use_relative = False

ARGS = {
    'predictions': predictions,
    'labels': labels,
    'mode_f1': mode_f1,
    'k_val': k_val,
    'use_relative': use_relative
}
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
div_backward Ran
ran; the contract check could not run (RAISES, DEP_MISSING or TIMEOUT)
yu-ni1989/SDCI/ovs/label_utils.py
code served (permissive licence) · get_code("941793a3319fabd3")
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, 3, 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)

# Analyze function signature and body:
# - px: shape (B, K, H, Wm1), used for indexing and arithmetic
# - py: shape (B, K, H-1, W), used for indexing and arithmetic
# The function creates div with shape (B, K, H, W) where W = Wm1 + 1

B = 2
K = 4
H = 3
Wm1 = 5
W = Wm1 + 1

# px has shape (B, K, H, Wm1)
px = torch.randn(B, K, H, Wm1, dtype=torch.float32)

# py has shape (B, K, H-1, W) - note H-1 for the middle dimension
py = torch.randn(B, K, H - 1, W, dtype=torch.float32)

ARGS = {
    "px": px,
    "py": py,
}
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, 3, 6) — shape (2, 4, 3, 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
drop_path Ran
ran; the contract check could not run (RAISES, DEP_MISSING or TIMEOUT)
yu-ni1989/SDCI/modified_clip/vision_transformer.py
pointer only (licence: MIT); the stored copy get_code serves, from gladzhang/Xformer, is pointer only (licence: NONE) · get_code("55120f2026b56aa2")
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 gladzhang/Xformer, 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.

felzenszwalb_img2labels Ran
ran, honoured the contract we drafted
yu-ni1989/SDCI/ovs/label_utils.py
code served (permissive licence) · get_code("1142943d7b9c8398")
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 int, shape (10, 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 numpy as np
from PIL import Image
import tempfile
import os

# Create a temporary RGB image file for testing
temp_dir = tempfile.gettempdir()
image_path = os.path.join(temp_dir, "test_image.png")

# Create a small RGB image (10x10x3)
img_array = np.random.randint(0, 256, (10, 10, 3), dtype=np.uint8)
img = Image.fromarray(img_array, mode='RGB')
img.save(image_path)

ARGS = {
    'image_path': image_path,
    'scale': 40,
    'sigma': 0.8,
    'min_size': 20
}
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 ndarray, ndim 2, dtype family int, shape (10, 10) — shape (10, 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
get_true_label Ran
ran; the contract check could not run (RAISES, DEP_MISSING or TIMEOUT)
yu-ni1989/SDCI/util/miou.py
code served (permissive licence) · get_code("d95287e5755c6662")
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 int, shape (64, 64) as a later run recorded it. The image it ran in was not recorded. If your copy returns the same type, ndim and shape on these inputs, it matched what we observed; nothing here checks that. It does NOT mean the output is correct: nothing has compared it to the paper.

import cv2
import os
import tempfile
import numpy as np

# Create a temporary directory structure for testing
temp_dir = tempfile.mkdtemp()
label_root = os.path.join(temp_dir, 'labels')
os.makedirs(label_root, exist_ok=True)

# Create a dummy label image file
# The function expects a file named like: <combined_path_str>_label.png
dummy_label_img = np.random.randint(0, 256, (64, 64), dtype=np.uint8)
label_filename = os.path.join(label_root, 'sample_data_label.png')
cv2.imwrite(label_filename, dummy_label_img)

# Define arguments
combined_path_str = 'sample_data_img'  # Will be converted to sample_data_label in the function
target_shape = (64, 64)  # A tuple representing target shape for the label

ARGS = {
    'combined_path_str': combined_path_str,
    'label_root': label_root,
    'target_shape': target_shape
}
Setup block drafted by
a language model, Claude Haiku 4.5, prompt python_fixture_v1, through the batch API
Imports
cv2, numpy — import names as written, not package names to install; a name the repository itself provides is not a package
Returned
type ndarray, ndim 2, dtype family int, shape (64, 64) — shape (64, 64) as a later run recorded it
dtype family
int means an integer dtype; float means any other dtype the instrument saw, bool and complex included (fixture_check.py records only these two families)
Environment
image not recorded; re-verify against syntology-harvest-sandbox:v5-deps; not recorded for this sample: the harness installs a cuda-to-cpu shim (cuda_shim.py) before every sample and no run records whether a sample needed it; seed: none by the harness; any seed is in the setup block
grad_forward Ran
ran; our contract draft was wrong, not the code
yu-ni1989/SDCI/ovs/label_utils.py
code served (permissive licence) · get_code("03bc00c3e031c484")
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. 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 numpy as np

torch.manual_seed(0)

# The function expects q to be a tensor that supports indexing with [...]
# It computes gradients along the last two dimensions
# q should be at least 2D for the slicing operations to work meaningfully
# Let's create a tensor with shape (2, 4, 5) - batch of 2, spatial 4x5
q = torch.randn(2, 4, 5)

ARGS = {
    'q': q
}
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 — a tuple 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
load_yaml Ran
ran; our contract draft was wrong, not the code
yu-ni1989/SDCI/util/tools.py
code served (permissive licence) · get_code("7ecf19d7f6016dce")
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 Config. 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 tempfile
import os
import yaml

# Create a temporary YAML file for testing
temp_dir = tempfile.mkdtemp()
yaml_file = os.path.join(temp_dir, "test_config.yaml")

# Write a sample YAML configuration
yaml_content = {
    "model": {
        "name": "test_model",
        "layers": 3,
        "hidden_dim": 64
    },
    "training": {
        "epochs": 10,
        "batch_size": 32,
        "learning_rate": 0.001
    },
    "dataset": "cifar10"
}

with open(yaml_file, 'w') as f:
    yaml.dump(yaml_content, f)

ARGS = {
    "filename": yaml_file
}
Setup block drafted by
a language model, Claude Haiku 4.5, prompt python_fixture_v1, through the batch API
Imports
yaml — import names as written, not package names to install; a name the repository itself provides is not a package
Returned
type Config — a Config output has no array shape of its own; none is stated — the instrument coerces any output to an array; for a Config 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
str2bool Ran
ran, violated the contract we drafted
yu-ni1989/SDCI/util/args.py
code served (permissive licence) · get_code("5aeffcaa5e40f745")
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 bool, 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)

ARGS = {
    'ipt': 'true'
}
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 bool, 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
trunc_normal_ Ran
ran; the contract check could not run (RAISES, DEP_MISSING or TIMEOUT)
yu-ni1989/SDCI/modified_clip/vision_transformer.py
code served (permissive licence) · get_code("fabe0899316ffaa5")
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 Tensor, ndim 2, dtype family float, shape (2, 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 torch
import math

torch.manual_seed(0)

# Create a small tensor to be initialized
tensor = torch.empty(2, 4)

# Define arguments for trunc_normal_
ARGS = {
    'tensor': tensor,
    'mean': 0.0,
    'std': 1.0,
    'a': -2.0,
    'b': 2.0
}
Setup block drafted by
a language model, Claude Haiku 4.5, prompt python_fixture_v1
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, 4) — shape (2, 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
vit_tiny Not yet run
unverified (harvested, no recorded run)
yu-ni1989/SDCI/modified_clip/vision_transformer.py
code served (permissive licence) · get_code("0b8fa7e1c9f0aa21")
How it ran

no invocation recorded.

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, 3 are also present in other papers whose code we hold (3 with at least one paper dated earlier, 2 the same month, 3 later) and 9 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
drop_path 224 other papers
217 dated earlier, 1 the same month, 5 later, 1 undated · 38 only as an assembly brick
2003.12039 (2020-03)
RAFT: Recurrent All-Pairs Field Transforms for Optical Flow
2007.09384 (2020-07)
Multi-Scale Positive Sample Refinement for Few-Shot Object Detection
2010.11929 (2020-10)
An Image is Worth 16x16 Words: Transformers for Image Recognition at Scale
and 221 more not named here
trunc_normal_ 160 other papers
150 dated earlier, 3 the same month, 7 later · 36 only as an assembly brick
1911.10194 (2019-11)
Panoptic-DeepLab: A Simple, Strong, and Fast Baseline for Bottom-Up Panoptic Segmentation
2005.10242 (2020-05)
Understanding Contrastive Representation Learning through Alignment and Uniformity on the Hypersphere
2007.12911 (2020-07)
Tighter risk certificates for neural networks
and 157 more not named here
vit_tiny 33 other papers
31 dated earlier, 1 later, 1 undated
2201.02609 (2022-01)
Generalized Category Discovery
2202.11929 (2022-02)
Word Segmentation on Discovered Phone Units with Dynamic Programming and Self-Supervised Scoring
2203.08414 (2022-03)
Unsupervised Semantic Segmentation by Distilling Feature Correspondences
and 30 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

High-resolution remote sensing images contain densely distributed objects with pronounced scale variations and complex boundaries, which impose higher demands on both the geometric localization and semantic prediction capabilities of semantic segmentation models. Existing training-free open-vocabulary semantic segmentation (OVSS) methods typically fuse Contrastive Language-Image Pretraining (CLIP) and vision foundation models (VFMs) using "oneway injection" and "shallow post-processing" strategies, making it difficult to satisfy these requirements. To address this issue, we propose a spatial-regularization-aware dual-branch collaborative inference framework for training-free OVSS, termed SDCI. First, during feature encoding, SDCI introduces a cross-model attention fusion (CAF) module, which guides collaborative inference by injecting self-attention maps into each other. Second, we propose a bidirectional cross-graph diffusion refinement (BCDR) module that enhances the reliability of dual-branch segmentation scores through iterative random-walk diffusion. Finally, we incorporate low-level superpixel structures and develop a convex-optimizationbased superpixel collaborative prediction (CSCP) mechanism to further refine object boundaries. Experiments on multiple remote sensing semantic segmentation benchmarks demonstrate that our method achieves better performance than existing approaches. Our code is available at https://github.com/yu-ni1989/SDCI. • Computing methodologies → Image segmentation; Neural networks; • Applied computing → Earth and atmospheric sciences.

For agents

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

get_harvested_code_for_paper("2601.21159")
get_code_for_paper("2601.21159")
have("2601.21159")

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

curl https://syntology.ai/api/ran/2601.21159.json

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

[![Syntology run record](https://syntology.ai/api/ran/2601.21159.svg)](https://syntology.ai/paper/2601.21159)

Connect an agent — have() is free.