SYNTOLOGY HomeExplorerAtlasCodeMethodologyAboutDevelopersFeedPricing
Paper · 2503.23925 · ICCV · 2025

CoMatch: Dynamic Covisibility-Aware Transformer for Bilateral Subpixel-Level Semi-Dense Image Matching

Yifan Lu, Jiayi Ma, Linfeng Tang, Shihua Zhang, Zizhuo Li, Token Condenser

arXiv · PDF · Open in the Atlas

Code that ran

2 of 4 samples (functions and classes) lifted from this paper's linked repositories executed on a synthesized input in Syntology's jail (1 of them constructed an object rather than computing a result; 1 honoured the contract we drafted, 0 violated it and 1 has no contract verdict; 0 failures of our instrument; 2 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/2503.23925.json · badge: /api/ran/2503.23925.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
ZizhuoLi/CoMatch community
Papers with Code archive
1 ran · 2 no recorded run
copy not recorded community
not in the paper's declared repositories
1 ran · 0 no recorded run
FunctionStatusWhere it lives
CoarseMatching Ran
ran on a synthesized input · constructed an object; no result was computed
ZizhuoLi/CoMatch/src/loftr/utils/coarse_matching.py
code served (permissive licence) · get_code("b5f7ae1e22cbbfd1")
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 an instance of its own class, CoarseMatching: the run constructed the object and did not call it, so no result was computed. 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 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)

# Create a minimal config dict with all required keys
config = {
    'thr': 0.1,
    'border_rm': 2,
    'skip_softmax': False,
    'fp16matmul': False,
    'train_coarse_percent': 0.2,
    'train_pad_num_gt_min': 200,
}

# Create sample tensors with small but distinct dimensions
batch_size = 2
num_features_0 = 12  # L dimension
num_features_1 = 10  # S dimension
feature_dim = 4      # C dimension

feat_c0 = torch.randn(batch_size, num_features_0, feature_dim)
feat_c1 = torch.randn(batch_size, num_features_1, feature_dim)

# Create mask tensors
mask_c0 = torch.ones(batch_size, num_features_0, dtype=torch.bool)
mask_c1 = torch.ones(batch_size, num_features_1, dtype=torch.bool)

# Create data dict with required keys
data = {
    'hw0_c': (3, 4),  # height and width of coarse feature 0
    'hw1_c': (3, 3),  # height and width of coarse feature 1
    'mask0': torch.ones(batch_size, 3, 4, dtype=torch.bool),
    'mask1': torch.ones(batch_size, 3, 3, dtype=torch.bool),
}

ARGS = {
    'config': config,
}
Setup block drafted by
a language model, Claude Haiku 4.5, prompt python_fixture_v1, through the batch API
Imports
einops, torch — import names as written, not package names to install; a name the repository itself provides is not a package
Returned
type CoarseMatching — the run constructed an instance of its own class and did not call it; no shape is stated — the instrument coerces any output to an array; for a CoarseMatching 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
compute_max_candidates Ran
ran, honoured the contract we drafted
this paper's copy was not recorded; identical code first harvested from yuansun-xjtu/icomma
pointer only; the stored copy get_code serves, from yuansun-xjtu/icomma, is pointer only (licence: NONE) · get_code("67a8b79fd250e22b")
How it ran

The body of this sample is withheld: this paper's copy of the code was not recorded (its link names no repository or source path), so no licence was read for it. 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.

mask_border Not yet run
unverified (harvested, no recorded run)
ZizhuoLi/CoMatch/src/loftr/utils/coarse_matching.py
code served (permissive licence) · get_code("6369920ec3b639d5")
How it ran

no invocation recorded.

mask_border_with_padding Not yet run
unverified (harvested, no recorded run)
ZizhuoLi/CoMatch/src/loftr/utils/coarse_matching.py
code served (permissive licence) · get_code("9ddd18b7ca3a5cd3")
How it ran

no invocation recorded.

Also present in other papers' code

Of 4 samples lifted from this paper's linked repositories, 4 carry a function identity and 0 do not. Of the 4 distinct functions with an identity, 1 are also present in other papers whose code we hold (1 with at least one paper dated earlier, 0 the same month, 0 later) and 3 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
compute_max_candidates
brick of this paper's assembly: helper
5 other papers
5 dated earlier
2104.00680 (2021-04)
LoFTR: Detector-Free Local Feature Matching with Transformers
2203.09645 (2022-03)
MatchFormer: Interleaving Attention in Transformers for Feature Matching
2204.10144 (2022-04)
A case for using rotation invariant features in state of the art feature matchers
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

This prospective study proposes CoMatch, a novel semidense image matcher with dynamic covisibility awareness and bilateral subpixel accuracy. Firstly, observing that modeling context interaction over the entire coarse feature map elicits highly redundant computation due to the neighboring representation similarity of tokens, a covisibilityguided token condenser is introduced to adaptively aggregate tokens in light of their covisibility scores that are dynamically estimated, thereby ensuring computational efficiency while improving the representational capacity of aggregated tokens simultaneously. Secondly, considering that feature interaction with non-covisible areas is distracting, which may degrade feature distinctiveness, a covisibilityassisted attention mechanism is deployed to selectively suppress irrelevant message broadcast from non-covisible reduced tokens, resulting in robust and compact attention to relevant rather than all ones. Thirdly, we find that at the fine-level stage, current methods adjust only the target view's keypoints to subpixel level, while those in the source view remain restricted at the coarse level and thus not informative enough, detrimental to keypoint location-sensitive usages. A simple yet potent fine correlation module is developed to refine matching candidates in both source and target views to subpixel level, attaining attractive performance improvement. Thorough experimentation across an array of public benchmarks affirms CoMatch's promising accuracy, efficiency, and generalizability. Code is available at https://github.com/ZizhuoLi/CoMatch.

For agents

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

get_harvested_code_for_paper("2503.23925")
get_code_for_paper("2503.23925")
have("2503.23925")

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

curl https://syntology.ai/api/ran/2503.23925.json

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

[![Syntology run record](https://syntology.ai/api/ran/2503.23925.svg)](https://syntology.ai/paper/2503.23925)

Connect an agent — have() is free.