SYNTOLOGY HomeExplorerAtlasCodeMethodologyAboutDevelopersFeedPricing
Paper · 2308.02097 · ICCV · 2023

Multi-interactive Feature Learning and a Full-time Multi-modality Benchmark for Image Fusion and Segmentation

Risheng Liu, Long Ma, Zhu Liu, Jinyuan Liu, Xin Fan, Wei Zhong, Guanyao Wu, Zhongxuan Luo

arXiv · PDF · Open in the Atlas

Code that ran

19 of 26 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; 3 honoured the contract we drafted, 1 violated it and 0 have no contract verdict; 15 failures of our instrument; 7 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/2308.02097.json · badge: /api/ran/2308.02097.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
jinyuanliu-cv/segmif canonical (classifier label)
classifier basis: census
12 ran · 1 no recorded run
runjia0124/coconet canonical (classifier label)
classifier basis: census
7 ran · 6 no recorded run
FunctionStatusWhere it lives
YCbCr2RGB Ran
ran, honoured the contract we drafted
runjia0124/coconet/utils/utils.py
code served (permissive licence) · get_code("92167228343b84a5")
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. 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 YCbCr tensors
# Y ranges from 16-235, Cb and Cr range from 16-240 in valid range
# Using small tensors with batch size 2 and spatial dimensions 4
Y = torch.randn(2, 4) * 50 + 128  # Values around 128
Cb = torch.randn(2, 4) * 50 + 128  # Values around 128
Cr = torch.randn(2, 4) * 50 + 128  # Values around 128

ARGS = {
    'Y': Y,
    'Cb': Cb,
    'Cr': Cr
}
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
YCrCb2RGB Ran
ran; the contract check could not run (RAISES, DEP_MISSING or TIMEOUT)
jinyuanliu-cv/segmif/core/model_fusion.py
pointer only (licence: NONE) · get_code("145edd13c3925a58")
How it ran

The body of this sample is withheld: GitHub detected no licence in this paper's repository, jinyuanliu-cv/segmif 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.

compute_cls_loss Ran
ran; the contract check could not run (RAISES, DEP_MISSING or TIMEOUT)
jinyuanliu-cv/segmif/core/loss.py
pointer only (licence: NONE) · get_code("c567234c5e7fa848")
How it ran

The body of this sample is withheld: GitHub detected no licence in this paper's repository, jinyuanliu-cv/segmif 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.

compute_cnt_loss Ran
ran; our contract draft was wrong, not the code
jinyuanliu-cv/segmif/core/loss.py
pointer only (licence: NONE) · get_code("2fe6b97c7e7cccbb")
How it ran

The body of this sample is withheld: GitHub detected no licence in this paper's repository, jinyuanliu-cv/segmif 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.

coords_fmap2orig Ran
ran; the contract check could not run (RAISES, DEP_MISSING or TIMEOUT)
jinyuanliu-cv/segmif/core/loss.py
pointer only (licence: NONE) · get_code("68ee569246bd7e3a")
How it ran

The body of this sample is withheld: GitHub detected no licence in this paper's repository, jinyuanliu-cv/segmif 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.

downsample Ran
ran; the contract check could not run (RAISES, DEP_MISSING or TIMEOUT)
jinyuanliu-cv/segmif/lap_loss.py
pointer only (licence: NONE) · get_code("1f6001e1c3b06556")
How it ran

The body of this sample is withheld: GitHub detected no licence in this paper's repository, jinyuanliu-cv/segmif 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.

fname_presuffix Ran
ran; our contract draft was wrong, not the code
runjia0124/coconet/utils/utils.py
code served (permissive licence) · get_code("71a890b56b7f51ea")
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 str. 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 os

# Define ARGS dictionary with all parameters for fname_presuffix
ARGS = {
    'fname': 'foo.nii.gz',
    'prefix': 'pre_',
    'suffix': '_post',
    'newpath': '/tmp',
    'use_ext': True
}
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 str — a str output has no array shape of its own; none is stated — the instrument coerces any output to an array; for a str 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
gauss_kernel Ran
ran, honoured the contract we drafted
jinyuanliu-cv/segmif/lap_loss.py
pointer only (licence: NONE) · get_code("4a150d48b91667f9")
How it ran

The body of this sample is withheld: GitHub detected no licence in this paper's repository, jinyuanliu-cv/segmif 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.

load_img_name_list Ran
ran; our contract draft was wrong, not the code
jinyuanliu-cv/segmif/datasets/voc_method.py
pointer only (licence: NONE) · get_code("0cc9285c14985e5d")
How it ran

The body of this sample is withheld: GitHub detected no licence in this paper's repository, jinyuanliu-cv/segmif 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.

normPRED Ran
ran, honoured the contract we drafted
runjia0124/coconet/utils/save_image.py
code served (permissive licence) · get_code("89e4412c100b0f9b")
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

torch.manual_seed(0)

# Create a small tensor with distinct dimensions
d = torch.randn(2, 4)

ARGS = {
    "d": d
}
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
normalize_img Ran
ran, violated the contract we drafted
jinyuanliu-cv/segmif/datasets/imutils.py
pointer only (licence: NONE) · get_code("d1a6979f946f9e74")
How it ran

The body of this sample is withheld: GitHub detected no licence in this paper's repository, jinyuanliu-cv/segmif 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.

normimage Ran
ran; our contract draft was wrong, not the code
runjia0124/coconet/utils/save_image.py
code served (permissive licence) · get_code("df54d7f9e9e4ab99")
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 int, shape (4, 5, 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 numpy as np

torch.manual_seed(0)

# Create a test tensor with shape (batch=1, channels=3, height=4, width=5)
# This matches the function's expectation: it accesses [0] to get first batch item
# and expects either shape[0] == 3 (3 channels) or shape[0] == 1 (1 channel)
test_tensor_3channel = torch.randn(1, 3, 4, 5)

ARGS = {
    "input_image": test_tensor_3channel,
    "save_cfg": True,
    "imtype": np.uint8,
    "isMask": False,
}
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 ndarray, ndim 3, dtype family int, shape (4, 5, 3) — shape (4, 5, 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
normimage_test Ran
ran; the contract check could not run (RAISES, DEP_MISSING or TIMEOUT)
runjia0124/coconet/utils/save_image.py
code served (permissive licence) · get_code("1d18356aa2fee4e0")
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 int, shape (4, 5, 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 numpy as np
from PIL import Image

torch.manual_seed(0)

# Test case 1: torch.Tensor input with 3 channels, save_cfg=True
input_tensor = torch.randn(1, 3, 4, 5)

# Test case 2: numpy array input
input_numpy = np.random.rand(4, 5, 3) * 255

ARGS = {
    "input_image": input_tensor,
    "save_cfg": True,
    "usebytescale": False,
    "imtype": np.uint8
}
Setup block drafted by
a language model, Claude Haiku 4.5, prompt python_fixture_v1, through the batch API
Imports
PIL, numpy, torch — 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 int, shape (4, 5, 3) — shape (4, 5, 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
prepare_data_path Ran
ran; our contract draft was wrong, not the code
jinyuanliu-cv/segmif/TaskFusion_dataset2.py
pointer only (licence: NONE); the stored copy get_code serves, from milliexie/fusionmamba, is pointer only (licence: NONE) · get_code("b79bae2be65a7ef0")
How it ran

The body of this sample is withheld: GitHub detected no licence in this paper's repository, jinyuanliu-cv/segmif 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.

random_scaling Ran
ran; the contract check could not run (RAISES, DEP_MISSING or TIMEOUT)
jinyuanliu-cv/segmif/datasets/imutils.py
pointer only (licence: NONE) · get_code("a920d698dd17a625")
How it ran

The body of this sample is withheld: GitHub detected no licence in this paper's repository, jinyuanliu-cv/segmif 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.

random_scaling2 Ran
ran; the contract check could not run (RAISES, DEP_MISSING or TIMEOUT)
jinyuanliu-cv/segmif/datasets/imutils.py
pointer only (licence: NONE) · get_code("49fc95c7acdb2ea4")
How it ran

The body of this sample is withheld: GitHub detected no licence in this paper's repository, jinyuanliu-cv/segmif 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.

split_filename Ran
ran; our contract draft was wrong, not the code
runjia0124/coconet/utils/utils.py
code served (permissive licence) · get_code("47482689d9a75517")
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 os

# The function split_filename takes a filename string and splits it into
# path, base filename, and extension. It handles special extensions like
# .nii.gz, .tar.gz, and .niml.dset

# Test with a simple filename path
ARGS = {
    'fname': '/home/data/subject.nii.gz'
}
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
upsample Ran
ran; the contract check could not run (RAISES, DEP_MISSING or TIMEOUT)
jinyuanliu-cv/segmif/lap_loss.py
pointer only (licence: NONE) · get_code("267b7d002ecac955")
How it ran

The body of this sample is withheld: GitHub detected no licence in this paper's repository, jinyuanliu-cv/segmif 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.

weights_to_cpu Ran
ran; our contract draft was wrong, not the code
runjia0124/coconet/utils/checkpoint.py
code served (permissive licence) · get_code("b683a315ccc25715")
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 OrderedDict. 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
from collections import OrderedDict

torch.manual_seed(0)

# Create a small state dict with GPU tensors (simulated on CPU)
# The function expects an OrderedDict with tensor values
state_dict = OrderedDict()
state_dict['layer1.weight'] = torch.randn(4, 3)
state_dict['layer1.bias'] = torch.randn(2)
state_dict['layer2.weight'] = torch.randn(5, 4)

ARGS = {
    'state_dict': state_dict
}
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 OrderedDict — a OrderedDict output has no array shape of its own; none is stated — the instrument coerces any output to an array; for a OrderedDict 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
RGB2YCrCb Not yet run
unverified (harvested, no recorded run)
jinyuanliu-cv/segmif/core/model_fusion.py
pointer only (licence: NONE) · get_code("6b1c47ea9127b33e")
How it ran

The body of this sample is withheld: GitHub detected no licence in this paper's repository, jinyuanliu-cv/segmif 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. No invocation recorded.

load_model Not yet run
unverified (harvested, no recorded run)
runjia0124/coconet/models/P_loss.py
code served (permissive licence) · get_code("bbfe8d248ce2ed69")
How it ran

no invocation recorded.

load_part Not yet run
unverified (harvested, no recorded run)
runjia0124/coconet/utils/checkpoint.py
code served (permissive licence) · get_code("0511b91fcab989b2")
How it ran

no invocation recorded.

measure_info Not yet run
unverified (harvested, no recorded run)
runjia0124/coconet/models/train_tasks.py
code served (permissive licence) · get_code("76e6bd620ca82ccc")
How it ran

no invocation recorded.

measure_module1 Not yet run
unverified (harvested, no recorded run)
runjia0124/coconet/models/train_tasks.py
code served (permissive licence) · get_code("4f4b727b0fc6cd95")
How it ran

no invocation recorded.

resume Not yet run
unverified (harvested, no recorded run)
runjia0124/coconet/utils/checkpoint.py
code served (permissive licence) · get_code("a3db2d0cc62c57ce")
How it ran

no invocation recorded.

softmax Not yet run
unverified (harvested, no recorded run)
runjia0124/coconet/models/train_tasks.py
code served (permissive licence) · get_code("7fbf8b3d61a9af4b")
How it ran

no invocation recorded.

Also present in other papers' code

Of 26 samples lifted from this paper's linked repositories, 26 carry a function identity and 0 do not. Of the 26 distinct functions with an identity, 9 are also present in other papers whose code we hold (2 with at least one paper dated earlier, 5 the same month, 5 later) and 17 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
downsample 4 other papers
1 dated earlier, 3 later
1603.08155 (2016-03)
Perceptual Losses for Real-Time Style Transfer and Super-Resolution
aaai_27987 (2024 (year only))
Optimize & Reduce: A Top-Down Approach for Image Vectorization
2406.09754 (2024-06)
LAVIB: A Large-scale Video Interpolation Benchmark
and 1 more not named here
normPRED 4 other papers
2 dated earlier, 2 later
2005.09007 (2020-05)
U$^2$-Net: Going Deeper with Nested U-Structure for Salient Object Detection
2101.04704 (2021-01)
Boundary-Aware Segmentation Network for Mobile and Web Applications
2411.11045 (2024-11)
StableV2V: Stablizing Shape Consistency in Video-to-Video Editing
and 1 more not named here
YCrCb2RGB 2 other papers
1 the same month, 1 later
2308.03979 (2023-08)
PAIF: Perception-Aware Infrared-Visible Image Fusion for Attack-Tolerant Semantic Segmentation
RGB2YCrCb 1 other paper
1 the same month
2308.03979 (2023-08)
PAIF: Perception-Aware Infrared-Visible Image Fusion for Attack-Tolerant Semantic Segmentation
compute_cls_loss 1 other paper
1 the same month
2308.03979 (2023-08)
PAIF: Perception-Aware Infrared-Visible Image Fusion for Attack-Tolerant Semantic Segmentation
compute_cnt_loss 1 other paper
1 the same month
2308.03979 (2023-08)
PAIF: Perception-Aware Infrared-Visible Image Fusion for Attack-Tolerant Semantic Segmentation
coords_fmap2orig 1 other paper
1 the same month
2308.03979 (2023-08)
PAIF: Perception-Aware Infrared-Visible Image Fusion for Attack-Tolerant Semantic Segmentation
prepare_data_path 1 other paper
1 later
2404.09498 (2024-04)
FusionMamba: Dynamic Feature Enhancement for Multimodal Image Fusion with Mamba
weights_to_cpu 1 other paper
1 later
05527 (2024 (year only))
A Semantic Space is Worth 256 Language Descriptions: Make Stronger Segmentation Models with Descriptive Properties

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

Figure 1: The chord diagram on the left shows the association and relative number of various labels in the proposed FMB dataset. Its branches display part of the well registered targets, which are extraordinarily affluent and manifold. The zoomedin regions on the right show the segmentation comparison in cyan circles and fusion comparison in orange circles. It is obvious that the proposed method is superior to SOTA methods on both visual effects and mIoU.

For agents

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

get_harvested_code_for_paper("2308.02097")
get_code_for_paper("2308.02097")
have("2308.02097")

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

curl https://syntology.ai/api/ran/2308.02097.json

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

[![Syntology run record](https://syntology.ai/api/ran/2308.02097.svg)](https://syntology.ai/paper/2308.02097)

Connect an agent — have() is free.