arxiv: 2604.08230 · v1 · submitted 2026-04-09 · 💻 cs.CV

Recognition: unknown

Generalization Under Scrutiny: Cross-Domain Detection Progresses, Pitfalls, and Persistent Challenges

Saniya M.Deshmukh , Kailash A. Hambarde , Hugo Proen\c{c}a

Authors on Pith no claims yet

Pith reviewed 2026-05-10 17:14 UTC · model grok-4.3

classification 💻 cs.CV

keywords cross-domain object detectiondomain adaptationobject detectiondomain shiftadaptation taxonomymulti-stage pipelinessurvey

0 comments

The pith

A taxonomy organizes cross-domain object detection methods by the adaptation paradigm and the pipeline stage they target.

A machine-rendered reading of the paper's core claim, the machinery that carries it, and where it could break.

The survey formulates cross-domain object detection as a multi-stage process in which domain shifts affect feature extraction, region proposals, and final classification in distinct ways. It then groups existing methods into categories according to whether they adapt at the image, feature, or output level and according to their modeling assumptions about the shift. This organization matters because simple transfer techniques that work for classification often fail for detection, leaving systems brittle when deployed across sensors, weather, or environments. The work also reviews standard datasets and evaluation practices while listing open problems that block more reliable detectors.

Core claim

Object detection under domain shift is inherently more complex than classification because domain variations propagate through every stage of the pipeline, and existing adaptation methods can be systematically categorized by the pipeline component they modify and by the assumptions they make about the nature of the shift.

What carries the argument

The conceptual taxonomy that sorts methods according to adaptation paradigms, modeling assumptions, and the specific detection-pipeline components they address.

If this is right

Adaptation at a single pipeline stage leaves the remaining stages exposed to domain shift.
Stage-specific analysis accounts for observed differences in how well various methods close the performance gap.
Benchmark suites should separately measure robustness at feature extraction, proposal, and classification stages.
Effective future systems will likely combine adaptations across multiple pipeline components rather than relying on one.

Where Pith is reading between the lines

These are editorial extensions of the paper, not claims the author makes directly.

The framework could be used to spot uncovered combinations of shift type and pipeline stage that have received little attention.
It suggests that progress may require detector architectures built from the start to support modular adaptation rather than retrofitting existing models.
Real-world settings such as autonomous driving may benefit from prioritizing adaptations for the most frequent shift types at the most sensitive stages.

Load-bearing premise

The taxonomy captures all important existing methods and the stage-wise description of how domain shift propagates accurately reflects the problem's structure.

What would settle it

A new adaptation technique that cannot be assigned to any category in the taxonomy, or a controlled experiment showing that domain-shift effects on detection accuracy do not follow the predicted pattern across pipeline stages.

read the original abstract

Object detection models trained on a source domain often exhibit significant performance degradation when deployed in unseen target domains, due to various kinds of variations, such as sensing conditions, environments and data distributions. Hence, regardless the recent breakthrough advances in deep learning-based detection technology, cross-domain object detection (CDOD) remains a critical research area. Moreover, the existing literature remains fragmented, lacking a unified perspective on the structural challenges underlying domain shift and the effectiveness of adaptation strategies. This survey provides a comprehensive and systematic analysis of CDOD. We start upon a problem formulation that highlights the multi-stage nature of object detection under domain shift. Then, we organize the existing methods through a conceptual taxonomy that categorizes approaches based on adaptation paradigms, modeling assumptions, and pipeline components. Furthermore, we analyze how domain shift propagates across detection stages and discuss why adaptation in object detection is inherently more complex than in classification. In addition, we review commonly used datasets, evaluation protocols, and benchmarking practices. Finally, we identify the key challenges and outline promising future research directions. Cohesively, this survey aims to provide a unified framework for understanding CDOD and to guide the development of more robust detection systems.

Editorial analysis

A structured set of objections, weighed in public.

Desk editor's note, referee report, simulated authors' rebuttal, and a circularity audit. Tearing a paper down is the easy half of reading it; the pith above is the substance, this is the friction.

Desk Editor's Note private letter to a colleague

This survey gives a practical taxonomy for cross-domain object detection and walks through stage-wise shift effects, but skips any account of how the papers were chosen.

read the letter

This survey pulls together work on cross-domain object detection and tries to give it some structure. The main things to note are the taxonomy that groups methods by adaptation paradigms and by which parts of the detection pipeline they address, plus the breakdown of how domain shift moves through stages like feature extraction and bounding box regression. That framing is the clearest part of the paper and explains why simple classification tricks often fail here. The review of datasets, evaluation protocols, and common benchmarking problems is also straightforward and points out real inconsistencies in how people measure progress. The future directions section flags issues like handling multiple simultaneous shifts, which match what practitioners run into in driving or surveillance settings. What the paper does well is start from a multi-stage problem formulation instead of treating detection as a black box. This makes the taxonomy feel grounded rather than arbitrary. The claim that adaptation is inherently harder than in classification follows logically from the stage analysis, even if it stays at the conceptual level. The soft spots are straightforward. The paper gives no details on the literature search—no databases, keywords, date range, or inclusion rules—so it is impossible to judge whether the taxonomy is complete or whether some lines of work were overlooked. The complexity argument would be stronger with even a small table summarizing reported performance drops across stages from the cited papers. Without that, readers have to take the increased-difficulty point on trust. This paper is for researchers who need a map of the CDOD area or who want to place their own adaptation method inside a larger picture. A reader coming from domain adaptation or from applied detection work will get the most out of the structure. It deserves a serious referee because the taxonomy could serve as a reference point for the community once the selection process is documented and the complexity discussion is backed by some numbers.

Referee Report

2 major / 2 minor

Summary. The manuscript surveys cross-domain object detection (CDOD). It formulates the problem to highlight the multi-stage nature of detectors under domain shift, organizes existing methods via a conceptual taxonomy based on adaptation paradigms, modeling assumptions, and pipeline components, analyzes how domain shift propagates across detection stages while arguing that adaptation is inherently more complex than in classification, reviews datasets and evaluation protocols, and identifies key challenges with future directions.

Significance. If the taxonomy is near-exhaustive and the stage-wise propagation analysis holds, the survey would supply a needed unified framework for a fragmented literature, helping researchers navigate adaptation strategies and focus on persistent robustness issues in CDOD. The explicit comparison of complexity to classification tasks could usefully steer future work away from direct transfer of classification techniques.

major comments (2)

[Abstract and Introduction] Abstract and Introduction: The central claim of a 'comprehensive and systematic analysis' is load-bearing for the taxonomy but rests on an unspecified literature review protocol (no databases, search strings, date range, or inclusion/exclusion criteria are described). Without this, the taxonomy's coverage cannot be verified and selection bias cannot be ruled out.
[Domain shift propagation analysis] Section on domain shift propagation and complexity comparison: The assertion that adaptation in object detection is inherently more complex than in classification is presented conceptually without quantitative backing, such as aggregated performance drops per stage or a meta-analysis of reported gaps in multi-stage pipelines versus classification baselines.

minor comments (2)

[Taxonomy section] Taxonomy figures or tables would benefit from explicit legends or color-coding that distinguishes adaptation paradigms from pipeline-component categories to improve readability.
[Datasets and benchmarks review] Ensure all cited datasets include at least one reference and a brief note on domain characteristics (e.g., synthetic vs. real) for quick reference.

Simulated Author's Rebuttal

2 responses · 0 unresolved

We thank the referee for the thoughtful and constructive comments on our survey paper. We address each major comment below and indicate the revisions we will make to strengthen the manuscript.

read point-by-point responses

Referee: [Abstract and Introduction] Abstract and Introduction: The central claim of a 'comprehensive and systematic analysis' is load-bearing for the taxonomy but rests on an unspecified literature review protocol (no databases, search strings, date range, or inclusion/exclusion criteria are described). Without this, the taxonomy's coverage cannot be verified and selection bias cannot be ruled out.

Authors: We agree with the referee that explicitly documenting the literature review protocol is essential for establishing the systematic nature of the survey and for allowing readers to assess potential selection bias. In the revised manuscript, we will add a new subsection in the Introduction (or a dedicated 'Survey Methodology' section) that details the search strategy. This will include the academic databases and repositories searched (e.g., Google Scholar, arXiv, IEEE Xplore), the specific keywords and Boolean combinations used, the publication date range considered, and the inclusion/exclusion criteria applied to select papers for the taxonomy. We believe this addition will directly address the concern and reinforce the credibility of our comprehensive analysis. revision: yes
Referee: [Domain shift propagation analysis] Section on domain shift propagation and complexity comparison: The assertion that adaptation in object detection is inherently more complex than in classification is presented conceptually without quantitative backing, such as aggregated performance drops per stage or a meta-analysis of reported gaps in multi-stage pipelines versus classification baselines.

Authors: The discussion on the inherent complexity of domain adaptation in object detection versus classification is grounded in the structural differences of the tasks: object detection involves multiple stages (feature extraction, region proposal, classification, and bounding box regression), each susceptible to domain shift, along with the need to handle both semantic and spatial variations. While we acknowledge that a quantitative meta-analysis could provide additional empirical support, the significant heterogeneity in experimental setups, datasets, and evaluation metrics across the CDOD literature makes aggregating performance drops without introducing confounding factors difficult. Nevertheless, to strengthen this section, we will partially revise it by incorporating specific quantitative examples drawn from representative papers in the survey, highlighting stage-wise performance degradations and comparative gaps relative to classification tasks. This will provide more concrete backing while maintaining the conceptual framework. revision: partial

Circularity Check

0 steps flagged

No circularity: survey organizes external literature without self-referential derivations or predictions

full rationale

This is a survey paper that formulates the CDOD problem, proposes a conceptual taxonomy of existing methods drawn from the broader literature, analyzes domain shift propagation conceptually, reviews datasets and protocols, and identifies challenges. No equations, fitted parameters, predictions, or derivations appear in the provided abstract or structure. The taxonomy and analysis rest on external works rather than reducing to self-citation chains or self-definitions. Literature selection criteria are not detailed, but this affects completeness rather than creating circularity by construction. The work is self-contained as a review against external benchmarks and receives the default non-circularity outcome.

Axiom & Free-Parameter Ledger

0 free parameters · 0 axioms · 0 invented entities

As a literature survey the paper introduces no new free parameters, mathematical axioms, or invented entities; it relies on standard concepts from computer vision and domain adaptation already established in the field.

pith-pipeline@v0.9.0 · 5523 in / 1036 out tokens · 37465 ms · 2026-05-10T17:14:19.730076+00:00 · methodology

discussion (0)

Reference graph

Works this paper leans on

117 extracted references · 102 canonical work pages

[1]

IEEE Transactions on Instrumentation and Measurement (2025) https://doi.org/10.1109/TIM.2025.3527619

Wang, Y., Qu, Z., Hu, Z., Yang, C., Huang, X., Zhao, Z., Zhai, Y.: Cross-domain multi-level feature adaptive alignment r-cnn for insulator defect detection in transmission lines. IEEE Transactions on Instrumentation and Measurement (2025) https://doi.org/10.1109/TIM.2025.3527619

work page doi:10.1109/tim.2025.3527619 2025
[2]

Journal of Visual Communication and Image Representation, 104534 (2025) https://doi.org/10.1016/j.jvcir.2025

Shi, Y., Guo, J., Wang, X., Wang, Y.: Tdenet: Three-branch distillation enhance- ment network for foggy scene object detection. Journal of Visual Communication and Image Representation, 104534 (2025) https://doi.org/10.1016/j.jvcir.2025. 104534

work page doi:10.1016/j.jvcir.2025 2025
[3]

In: Proceedings of the IEEE/CVF International Conference on Computer Vision, pp

Liang, A., Kong, L., Lu, D., Liu, Y., Fang, J., Zhao, H., Ooi, W.T.: Perspective- invariant 3d object detection. In: Proceedings of the IEEE/CVF International Conference on Computer Vision, pp. 27725–27738 (2025)

2025
[4]

In: CVPR (2018)

Chen, Y., Li, W., Sakaridis, C., Dai, D., Van Gool, L.: Domain adaptive faster r- cnn for object detection in the wild. In: CVPR (2018). https://doi.org/10.1109/ cvpr.2018.00352

work page arXiv 2018
[5]

In: Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition, pp

Zhu, X., Pang, J., Yang, C., Shi, J., Lin, D.: Adapting object detectors via selective cross-domain alignment. In: Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition, pp. 687–696 (2019). https://doi. org/10.1109/cvpr.2019.00078

work page doi:10.1109/cvpr.2019.00078 2019
[6]

MViTv2: Improved Multiscale Vision Transformers for Classification and Detection , isbn =

Zhao, L., Wang, L.: Task-specific inconsistency alignment for domain adaptive object detection. In: Proceedings of the IEEE/CVF Conference on Computer 32 Vision and Pattern Recognition, pp. 14217–14226 (2022). https://doi.org/10. 1109/cvpr52688.2022.01382

work page arXiv 2022
[7]

IEEE Transactions on Pattern Analysis and Machine Intelligence46(3), 1742–1756 (2022) https://doi

Zhang, H., Xiao, L., Cao, X., Foroosh, H.: Multiple adverse weather conditions adaptation for object detection via causal intervention. IEEE Transactions on Pattern Analysis and Machine Intelligence46(3), 1742–1756 (2022) https://doi. org/10.1109/tpami.2022.3166765

work page doi:10.1109/tpami.2022.3166765 2022
[8]

In: CVPR (2019)

Saito, K., Watanabe, Y., Ushiku, Y., Harada, T.: Strong-weak distribution align- ment for adaptive object detection. In: CVPR (2019). https://doi.org/10.1109/ cvpr.2019.00712

work page arXiv 2019
[9]

In: CVPR

Liu, Y., Zhou, S., Liu, X., Hao, C., Fan, B., Tian, J.: Unbiased faster r- cnn for single-source domain generalized object detection. In: Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition, pp. 28838–28847 (2024). https://doi.org/10.1109/cvpr52733.2024.02724

work page doi:10.1109/cvpr52733.2024.02724 2024
[10]

IEEE Transactions on knowl- edge and data engineering22(10), 1345–1359 (2009) https://doi.org/10.1109/ TKDE.2009.191

Pan, S.J., Yang, Q.: A survey on transfer learning. IEEE Transactions on knowl- edge and data engineering22(10), 1345–1359 (2009) https://doi.org/10.1109/ TKDE.2009.191

2009
[11]

Liuyi Yao, Sheng Li, Yaliang Li, Mengdi Huai, Jing Gao, and Aidong Zhang

Weiss, K., Khoshgoftaar, T.M., Wang, D.: A survey of transfer learning. Journal of Big data3(1), 9 (2016) https://doi.org/10.1186/s40537-016-0043-6

work page doi:10.1186/s40537-016-0043-6 2016
[12]

arXiv preprint arXiv:1702.05374 (2017) https://doi.org/10.1007/ 978-3-319-58347-1 1

Csurka, G.: Domain adaptation for visual applications: A comprehen- sive survey. arXiv preprint arXiv:1702.05374 (2017) https://doi.org/10.1007/ 978-3-319-58347-1 1

work page arXiv 2017
[13]

Sim-to-Real Transfer in Deep Reinforcement Learning for Robotics: a Survey

Li, W., Li, F., Luo, Y., Wang, P.,et al.: Deep domain adaptive object detec- tion: A survey. In: 2020 IEEE Symposium Series on Computational Intelligence (SSCI), pp. 1808–1813 (2020). https://doi.org/10.1109/ssci47803.2020.9308604 . IEEE

work page doi:10.1109/ssci47803.2020.9308604 2020
[14]

arXiv preprint arXiv:2107.07927 (2021)

Muzammul, M., Li, X.: A survey on deep domain adaptation and tiny object detection challenges, techniques and datasets. arXiv preprint arXiv:2107.07927 (2021)

work page arXiv 2021
[15]

ACDC: The adverse conditions dataset with correspondences for robust semantic driving scene perception,

Oza, P., Sindagi, V.A., Patel, V.M.,et al.: Unsupervised domain adaptation of object detectors: A survey. IEEE Transactions on Pattern Analysis and Machine Intelligence46(6), 4018–4040 (2023) https://doi.org/10.1109/tpami. 2022.3217046

work page doi:10.1109/tpami 2023
[16]

Deep learning with edge computing: A review,

Zou, Z., Chen, K., Shi, Z., Guo, Y., Ye, J.: Object detection in 20 years: A survey. Proceedings of the IEEE111(3), 257–276 (2023) https://doi.org/10.1109/jproc. 2023.3238524 33

work page doi:10.1109/jproc 2023
[17]

ACM Computing Surveys57(8), 1–37 (2025) https://doi.org/10.1145/3718362

Xu, H., Zhi, S., Sun, S., Patel, V., Liu, L.: Deep learning for cross-domain few- shot visual recognition: A survey. ACM Computing Surveys57(8), 1–37 (2025) https://doi.org/10.1145/3718362

work page doi:10.1145/3718362 2025
[18]

In: Proceedings of the AAAI Conference on Artificial Intelligence, vol

Zheng, Y., Wu, J., Li, W., Chen, Z.: Universal domain adaptive object detection via dual probabilistic alignment. In: Proceedings of the AAAI Conference on Artificial Intelligence, vol. 39, pp. 10644–10652 (2025). https://doi.org/10.1609/ aaai.v39i10.33156

2025
[19]

Light-weight Calibrator:

Pan, Y., Yao, T., Li, Y., Ngo, C.-W., Mei, T.: Exploring category-agnostic clusters for open-set domain adaptation. In: Proceedings of the IEEE/CVF Con- ference on Computer Vision and Pattern Recognition, pp. 13867–13875 (2020). https://doi.org/10.1109/cvpr42600.2020.01388

work page doi:10.1109/cvpr42600.2020.01388 2020
[20]

Learning to count everything, in: IEEE Conference on Computer Vision and Pat- tern Recognition, CVPR 2021, virtual, June 19-25, 2021, Computer Vision Foundation / IEEE

Vs, V., Gupta, V., Oza, P., Sindagi, V.A., Patel, V.M.: Mega-cda: Memory guided attention for category-aware unsupervised domain adaptive object detec- tion. In: Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition, pp. 4516–4526 (2021). https://doi.org/10.1109/cvpr46437. 2021.00449

work page doi:10.1109/cvpr46437 2021
[21]

MViTv2: Improved Multiscale Vision Transformers for Classification and Detection , isbn =

Huang, J., Guan, D., Xiao, A., Lu, S., Shao, L.: Category contrast for unsu- pervised domain adaptation in visual tasks. In: Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition, pp. 1203–1214 (2022). https://doi.org/10.1109/cvpr52688.2022.00127

work page doi:10.1109/cvpr52688.2022.00127 2022
[22]

International Journal of Computer Vision129(7), 2223–2243 (2021) https://doi.org/10.1007/s11263-021-01447-x

Chen, Y., Wang, H., Li, W., Sakaridis, C., Dai, D., Van Gool, L.: Scale-aware domain adaptive faster r-cnn. International Journal of Computer Vision129(7), 2223–2243 (2021) https://doi.org/10.1007/s11263-021-01447-x

work page doi:10.1007/s11263-021-01447-x 2021
[23]

Multimedia Systems31(1), 24 (2025) https://doi.org/10.1007/s00530-024-01594-4

Wang, H., Qian, H.: Sr-dayolov8: cross-domain adaptive object detection based on super-resolution domain classifier. Multimedia Systems31(1), 24 (2025) https://doi.org/10.1007/s00530-024-01594-4

work page doi:10.1007/s00530-024-01594-4 2025
[24]

IEEE Transactions on Image Processing (2025) https://doi.org/10.1109/tip.2024.3522807

Chen, J., Liu, L., Deng, W., Liu, Z., Liu, Y., Wei, Y., Liu, Y.: Refining pseudo labeling via multi-granularity confidence alignment for unsupervised cross domain object detection. IEEE Transactions on Image Processing (2025) https://doi.org/10.1109/tip.2024.3522807

work page doi:10.1109/tip.2024.3522807 2025
[25]

Wille, C

Chen, L., Song, S., Wang, Y., Hu, Y., Han, J.: Gaussian-driven unsupervised domain adaptation object detection transformer for remote sensing imagery. IEEE Journal of Selected Topics in Applied Earth Observations and Remote Sensing (2025) https://doi.org/10.1109/jstars.2025.3608554

work page doi:10.1109/jstars.2025.3608554 2025
[26]

IEEE Transactions on Intelligent Transportation Systems25(11), 15977–15990 (2024) https://doi.org/10.1109/tits.2024.3413813 34

Cai, M., Kezierbieke, J., Zhong, X., Chen, H.: Uncertainty-aware and class- balanced domain adaptation for object detection in driving scenes. IEEE Transactions on Intelligent Transportation Systems25(11), 15977–15990 (2024) https://doi.org/10.1109/tits.2024.3413813 34

work page doi:10.1109/tits.2024.3413813 2024
[27]

In: Proceedings of the 28th ACM International Conference on Multimedia, pp

Nguyen, D.-K., Tseng, W.-L., Shuai, H.-H.: Domain-adaptive object detection via uncertainty-aware distribution alignment. In: Proceedings of the 28th ACM International Conference on Multimedia, pp. 2499–2507 (2020). https://doi.org/ 10.1145/3394171.3413553

work page doi:10.1145/3394171.3413553 2020
[28]

IEEE Transactions on Image Processing (2025) https://doi.org/10.1109/TIP.2025

Yao, H., Zhao, S., Lu, S., Chen, H., Li, Y., Liu, G., Xing, T., Yan, C., Tao, J., Ding, G.: Source-free object detection with detection transformer. IEEE Transactions on Image Processing (2025) https://doi.org/10.1109/TIP.2025. 3607621

work page doi:10.1109/tip.2025 2025
[29]

In: European Con- ference on Computer Vision, pp

Diamant, I., Rosenfeld, A., Achituve, I., Goldberger, J., Netzer, A.: De- confusing pseudo-labels in source-free domain adaptation. In: European Con- ference on Computer Vision, pp. 108–125 (2024). https://doi.org/10.1007/ 978-3-031-72986-7 7 . Springer

2024
[30]

2019.01.103

Jiang, W., Luan, Y., Tang, K., Wang, L., Zhang, N., Chen, H., Qi, H.: Adaptive feature alignment network with noise suppression for cross-domain object detec- tion. Neurocomputing614, 128789 (2025) https://doi.org/10.1016/j.neucom. 2024.128789

work page doi:10.1016/j.neucom 2025
[31]

In: Proceedings of the AAAI Conference on Artificial Intelligence, vol

He, X., Li, X., Guo, X.: Differential alignment for domain adaptive object detec- tion. In: Proceedings of the AAAI Conference on Artificial Intelligence, vol. 39, pp. 17150–17158 (2025). https://doi.org/10.1609/aaai.v39i16.33885

work page doi:10.1609/aaai.v39i16.33885 2025
[32]

Advances in Neu- ral Information Processing Systems36, 4248–4262 (2023) https://doi.org/10

Li, H., Zhang, R., Yao, H., Song, X., Hao, Y., Zhao, Y., Li, L., Chen, Y.: Learning domain-aware detection head with prompt tuning. Advances in Neu- ral Information Processing Systems36, 4248–4262 (2023) https://doi.org/10. 52202/075280-0187

2023
[33]

In: CVPR

Zhang, Z., Chen, M., Xiao, S., Peng, L., Li, H., Lin, B., Li, P., Wang, W., Wu, B., Cai, D.: Pseudo label refinery for unsupervised domain adaptation on cross-dataset 3d object detection. In: Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition, pp. 15291–15300 (2024). https: //doi.org/10.1109/cvpr52733.2024.01448

work page doi:10.1109/cvpr52733.2024.01448 2024
[34]

MViTv2: Improved Multiscale Vision Transformers for Classification and Detection , isbn =

Xu, Y., Sun, Y., Yang, Z., Miao, J., Yang, Y.: H2fa r-cnn: Holistic and hier- archical feature alignment for cross-domain weakly supervised object detection. In: Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition, pp. 14329–14339 (2022). https://doi.org/10.1109/cvpr52688.2022. 01393

work page doi:10.1109/cvpr52688.2022 2022
[35]

Pattern Recognition158, 111024 (2025) https://doi.org/10.1016/j.patcog.2024.111024

Yang, R., Tian, T., Tian, J.: Versatile teacher: A class-aware teacher–student framework for cross-domain adaptation. Pattern Recognition158, 111024 (2025) https://doi.org/10.1016/j.patcog.2024.111024

work page doi:10.1016/j.patcog.2024.111024 2025
[36]

Light-weight Calibrator:

Zheng, Y., Huang, D., Liu, S., Wang, Y.: Cross-domain object detection through coarse-to-fine feature adaptation. In: Proceedings of the IEEE/CVF Conference 35 on Computer Vision and Pattern Recognition, pp. 13766–13775 (2020). https: //doi.org/10.1109/cvpr42600.2020.01378

work page doi:10.1109/cvpr42600.2020.01378 2020
[37]

Light-weight Calibrator:

Xu, M., Wang, H., Ni, B., Tian, Q., Zhang, W.: Cross-domain detection via graph-induced prototype alignment. In: Proceedings of the IEEE/CVF Confer- ence on Computer Vision and Pattern Recognition, pp. 12355–12364 (2020). https://doi.org/10.1109/cvpr42600.2020.01237

work page doi:10.1109/cvpr42600.2020.01237 2020
[38]

In: Proceedings of the AAAI Conference on Artificial Intelligence, vol

Li, W., Liu, X., Yao, X., Yuan, Y.: Scan: Cross domain object detection with semantic conditioned adaptation. In: Proceedings of the AAAI Conference on Artificial Intelligence, vol. 36, pp. 1421–1428 (2022). https://doi.org/10.1609/ aaai.v36i2.20031

2022
[39]

2021 , url =

Deng, J., Li, W., Chen, Y., Duan, L.: Unbiased mean teacher for cross-domain object detection. In: Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition, pp. 4091–4101 (2021). https://doi.org/10.1109/ cvpr46437.2021.00408

work page arXiv 2021
[40]

IEEE Transactions on Image Processing30, 4046–4056 (2021) https://doi.org/10.1109/tip.2021.3066046

Wang, H., Liao, S., Shao, L.: Afan: Augmented feature alignment network for cross-domain object detection. IEEE Transactions on Image Processing30, 4046–4056 (2021) https://doi.org/10.1109/tip.2021.3066046

work page doi:10.1109/tip.2021.3066046 2021
[41]

Expert Systems with Applications205, 117697 (2022) https://doi.org/10.2139/ ssrn.4062473

Do, M., Jeon, S., Lee, P., Hong, K., Ma, Y.-s., Byun, H.: Exploiting domain transferability for collaborative inter-level domain adaptive object detection. Expert Systems with Applications205, 117697 (2022) https://doi.org/10.2139/ ssrn.4062473

2022
[42]

Neural Computing and Appli- cations36(7), 3631–3644 (2024) https://doi.org/10.1007/s00521-023-09248-8

Song, Y., Liu, Z., Tang, R., Duan, G., Tan, J.: Cross-domain object detection by local to global object-aware feature alignment. Neural Computing and Appli- cations36(7), 3631–3644 (2024) https://doi.org/10.1007/s00521-023-09248-8

work page doi:10.1007/s00521-023-09248-8 2024
[43]

IEEE Transactions on Neural Networks and Learning Systems35(11), 15170–15181 (2023) https://doi.org/10.1109/tnnls

Piao, Z., Tang, L., Zhao, B.: Unsupervised domain-adaptive object detection via localization regression alignment. IEEE Transactions on Neural Networks and Learning Systems35(11), 15170–15181 (2023) https://doi.org/10.1109/tnnls. 2023.3282958

work page doi:10.1109/tnnls 2023
[44]

arXiv preprint arXiv:2403.12029 (2024)

Kay, J., Haucke, T., Stathatos, S., Deng, S., Young, E., Perona, P., Beery, S., Van Horn, G.: Align and distill: Unifying and improving domain adaptive object detection. arXiv preprint arXiv:2403.12029 (2024)

work page arXiv 2024
[45]

Khan, and Fahad Shah- baz Khan

Wu, A., Liu, R., Han, Y., Zhu, L., Yang, Y.: Vector-decomposed disentangle- ment for domain-invariant object detection. In: Proceedings of the IEEE/CVF International Conference on Computer Vision, pp. 9342–9351 (2021). https: //doi.org/10.1109/iccv48922.2021.00921

work page doi:10.1109/iccv48922.2021.00921 2021
[46]

Advances in neural information 36 processing systems32(2019)

Zhang, Q., Zhang, J., Liu, W., Tao, D.: Category anchor-guided unsupervised domain adaptation for semantic segmentation. Advances in neural information 36 processing systems32(2019)

2019
[47]

Neurocomputing440, 310–320 (2021) https://doi.org/10.1016/j.neucom.2021

Iqbal, J., Munir, M.A., Mahmood, A., Ali, A.R., Ali, M.: Leveraging orientation for weakly supervised object detection with application to firearm localization. Neurocomputing440, 310–320 (2021) https://doi.org/10.1016/j.neucom.2021. 01.075

work page doi:10.1016/j.neucom.2021 2021
[48]

Engineering Applications of Artificial Intelligence 157, 111014 (2025) https://doi.org/10.1016/j.engappai.2025.111014

Wei, X., Xia, J., Yang, F., Zhao, C., Liu, C., Wang, G., Chen, Y., Lu, Y.: Multi- scale pseudo-labels filtering and key pixels adversarial alignment for domain adaptive object detection. Engineering Applications of Artificial Intelligence 157, 111014 (2025) https://doi.org/10.1016/j.engappai.2025.111014

work page doi:10.1016/j.engappai.2025.111014 2025
[49]

Revisiting pre-trained remote sensing model benchmarks: Resizing and normalization matters

Kim, J., Ku, Y., Kim, J., Cha, J., Baek, S.: Vlm-pl: Advanced pseudo labeling approach for class incremental object detection via vision-language model. In: Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition, pp. 4170–4181 (2024). https://doi.org/10.1109/cvprw63382.2024. 00420

work page doi:10.1109/cvprw63382.2024 2024
[50]

Sensors25(11), 3433 (2025) https://doi.org/10

Luo, Y., Wu, A., Fu, Q.: Mas-yolov11: An improved underwater object detection algorithm based on yolov11. Sensors25(11), 3433 (2025) https://doi.org/10. 3390/s25113433

2025
[51]

Information Fusion117, 102871 (2025) https://doi.org/10.1016/j.inffus.2024.102871

Wang, T., Yu, Z., Fang, J., Xie, J., Yang, F., Zhang, H., Zhang, L., Du, M., Li, L., Ning, X.: Multidimensional fusion of frequency and spatial domain information for enhanced camouflaged object detection. Information Fusion117, 102871 (2025) https://doi.org/10.1016/j.inffus.2024.102871

work page doi:10.1016/j.inffus.2024.102871 2025
[52]

In: Proceedings of the Computer Vision and Pattern Recognition Conference, pp

Lavoie, M.-A., Mahmoud, A., Waslander, S.L.: Large self-supervised models bridge the gap in domain adaptive object detection. In: Proceedings of the Computer Vision and Pattern Recognition Conference, pp. 4692–4702 (2025)

2025
[53]

IEEE Transactions on Geoscience and Remote Sensing (2025) https://doi.org/10

Yang, B., Han, J., Hou, X., Zhou, D., Liu, W., Bi, F.: Fsda-detr: Few-shot domain adaptive object detection transformer in remote sensing imagery. IEEE Transactions on Geoscience and Remote Sensing (2025) https://doi.org/10. 1109/tgrs.2025.3574245

work page arXiv 2025
[54]

Ieee Access8, 182105–182116 (2020) https://doi.org/10.1109/access.2020

Zhang, P., Zhang, Z., Hao, Y., Zhou, Z., Luo, B., Wang, T.: Multi-scale feature enhanced domain adaptive object detection for power transmission line inspec- tion. Ieee Access8, 182105–182116 (2020) https://doi.org/10.1109/access.2020. 3027850

work page doi:10.1109/access.2020 2020
[55]

Electronics13(10), 1823 (2024) https://doi.org/10.3390/electronics13101823

Zhou, R., Wang, Q., Cao, L., Xu, J., Zhu, X., Xiong, X., Zhang, H., Zhong, Y.: Dual-level viewpoint-learning for cross-domain vehicle re-identification. Electronics13(10), 1823 (2024) https://doi.org/10.3390/electronics13101823

work page doi:10.3390/electronics13101823 2024
[56]

arXiv preprint arXiv:1911.07158 (2019)

Yu, F., Wang, D., Chen, Y., Karianakis, N., Shen, T., Yu, P., Lymberopoulos, 37 D., Lu, S., Shi, W., Chen, X.: Unsupervised domain adaptation for object detec- tion via cross-domain semi-supervised learning. arXiv preprint arXiv:1911.07158 (2019)

work page arXiv 1911
[57]

Science of Remote Sensing11, 100202 (2025) https://doi.org/10.2139/ ssrn.5049770

Zhao, S., Kang, Y., Yuan, H., Wang, G., Wang, H., Xiong, S., Luo, Y.: Fsdaod: Few-shot domain adaptation object detection for heterogeneous sar image. Science of Remote Sensing11, 100202 (2025) https://doi.org/10.2139/ ssrn.5049770

2025
[58]

Leveraging vision language models for specialized agricultural tasks

Shangguan, Z., Seita, D., Rostami, M.: Cross-domain multi-modal few-shot object detection via rich text. In: 2025 IEEE/CVF Winter Conference on Appli- cations of Computer Vision (WACV), pp. 6570–6580 (2025). https://doi.org/10. 1109/wacv61041.2025.00640 . IEEE

work page arXiv 2025
[59]

Ben-David, J

Ben-David, S., Blitzer, J., Crammer, K., Kulesza, A., Pereira, F., Vaughan, J.W.: A theory of learning from different domains. Machine learning79(1), 151–175 (2010) https://doi.org/10.1007/s10994-009-5152-4

work page doi:10.1007/s10994-009-5152-4 2010
[60]

IEEE Transactions on Intelligent Transportation Systems23(8), 12633–12647 (2021) https://doi.org/10.1109/tits.2021.3115823

Zhang, H., Luo, G., Li, J., Wang, F.-Y.: C2fda: Coarse-to-fine domain adaptation for traffic object detection. IEEE Transactions on Intelligent Transportation Systems23(8), 12633–12647 (2021) https://doi.org/10.1109/tits.2021.3115823

work page doi:10.1109/tits.2021.3115823 2021
[61]

Remote Sensing16(5), 907 (2024) https://doi.org/10.3390/rs16050907

Liu, C., Zhang, S., Hu, M., Song, Q.: Object detection in remote sensing images based on adaptive multi-scale feature fusion method. Remote Sensing16(5), 907 (2024) https://doi.org/10.3390/rs16050907

work page doi:10.3390/rs16050907 2024
[62]

IEEE Transactions on Image Processing34, 729–742 (2024) https: //doi.org/10.1109/tip.2024.3459589

He, Y., Chen, W., Wang, S., Liu, T., Wang, M.: Recalling unknowns without losing precision: An effective solution to large model-guided open world object detection. IEEE Transactions on Image Processing34, 729–742 (2024) https: //doi.org/10.1109/tip.2024.3459589

work page doi:10.1109/tip.2024.3459589 2024
[63]

In: Proceed- ings of the IEEE Conference on Computer Vision and Pattern Recognition, pp

Inoue, N., Furuta, R., Yamasaki, T., Aizawa, K.: Cross-domain weakly- supervised object detection through progressive domain adaptation. In: Proceed- ings of the IEEE Conference on Computer Vision and Pattern Recognition, pp. 5001–5009 (2018). https://doi.org/10.1109/cvpr.2018.00525

work page doi:10.1109/cvpr.2018.00525 2018
[64]

In: Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition, pp

Liu, X., Wang, Z., Shao, J., Wang, X., Li, H.: Improving referring expres- sion grounding with cross-modal attention-guided erasing. In: Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition, pp. 1950–1959 (2019). https://doi.org/10.1109/cvpr.2019.00205

work page doi:10.1109/cvpr.2019.00205 1950
[65]

IEEE Transactions on Geoscience and Remote Sensing62, 1–15 (2024) https://doi.org/10.36227/techrxiv.24745587

Biswas, D., Teˇ si´ c, J.: Domain adaptation with contrastive learning for object detection in satellite imagery. IEEE Transactions on Geoscience and Remote Sensing62, 1–15 (2024) https://doi.org/10.36227/techrxiv.24745587

work page doi:10.36227/techrxiv.24745587 2024
[66]

IET 38 Image Processing20(1), 70294 (2026) https://doi.org/10.1049/ipr2.70294

Geng, H., Fang, L., Wang, Y., Liu, Z., Fan, Z.: Cen-rtdetr: A co-enhancement- based real-time single-domain generalized object detection for road scenes. IET 38 Image Processing20(1), 70294 (2026) https://doi.org/10.1049/ipr2.70294

work page doi:10.1049/ipr2.70294 2026
[67]

arXiv preprint arXiv:2312.15275 (2023)

Saoud, L.S., Seneviratne, L., Hussain, I.: Mars: Multi-scale adaptive robotics vision for underwater object detection and domain generalization. arXiv preprint arXiv:2312.15275 (2023)

work page arXiv 2023
[68]

In: 2024 IEEE International Conference on Image Processing (ICIP), pp

Saoud, L.S., Niu, Z., Seneviratne, L., Hussain, I.: Real-time and resource-efficient multi-scale adaptive robotics vision for underwater object detection and domain generalization. In: 2024 IEEE International Conference on Image Processing (ICIP), pp. 3917–3923 (2024). https://doi.org/10.1109/icip51287.2024.10647684 . IEEE

work page doi:10.1109/icip51287.2024.10647684 2024
[69]

In: 2025 IEEE International Conference on Advanced Visual and Signal-Based Systems (AVSS), pp

Tulu, A.W., Conci, N.: Wct-enhanced instance normalization for unsupervised domain adaptation in object detection. In: 2025 IEEE International Conference on Advanced Visual and Signal-Based Systems (AVSS), pp. 1–6 (2025). https: //doi.org/10.1109/avss65446.2025.11149950 . IEEE

work page doi:10.1109/avss65446.2025.11149950 2025
[70]

IEEE Transactions on Circuits and Systems for Video Technology (2024) https://doi.org/10.1109/ tcsvt.2024.3520734

Xu, S., Li, X., Wu, S., Zhang, W., Tong, Y., Loy, C.C.: Dst-det: Open- vocabulary object detection via dynamic self-training. IEEE Transactions on Circuits and Systems for Video Technology (2024) https://doi.org/10.1109/ tcsvt.2024.3520734

work page arXiv 2024
[71]

IEEE Transactions on Geoscience and Remote Sensing60, 1–11 (2022) https://doi.org/10.1109/tgrs.2022.3183022

Cheng, G., Wang, J., Li, K., Xie, X., Lang, C., Yao, Y., Han, J.: Anchor-free ori- ented proposal generator for object detection. IEEE Transactions on Geoscience and Remote Sensing60, 1–11 (2022) https://doi.org/10.1109/tgrs.2022.3183022

work page doi:10.1109/tgrs.2022.3183022 2022
[72]

In: Pro- ceedings of the 7th ACM International Conference on Multimedia in Asia, pp

Zhou, L., Wang, R., Xue, L., Yang, J.: Ccanet: A cognition-inspired framework for few-shot segmentation from category-agnostic to category-aware. In: Pro- ceedings of the 7th ACM International Conference on Multimedia in Asia, pp. 1–8 (2025). https://doi.org/10.1145/3743093.3770980

work page doi:10.1145/3743093.3770980 2025
[73]

arXiv preprint arXiv:2311.17942 (2023)

Niu, D., Bar, A., Herzig, R., Darrell, T., Rohrbach, A.: Object-based (yet class- agnostic) video domain adaptation. arXiv preprint arXiv:2311.17942 (2023)

work page arXiv 2023
[74]

Bridging Annotation Gaps: Transferring Labels to Align Object Detection Datasets,

Kennerley, M., Aviles-Rivero, A., Sch¨ onlieb, C.-B., Tan, R.T.: Bridging annota- tion gaps: Transferring labels to align object detection datasets. arXiv preprint arXiv:2506.04737 (2025)

work page arXiv 2025
[75]

IEEE Transactions on Pattern Analysis and Machine Intelli- gence45(6), 7338–7352 (2022) https://doi.org/10.1109/tpami.2022.3218569

Jiao, Y., Yao, H., Xu, C.: Dual instance-consistent network for cross-domain object detection. IEEE Transactions on Pattern Analysis and Machine Intelli- gence45(6), 7338–7352 (2022) https://doi.org/10.1109/tpami.2022.3218569

work page doi:10.1109/tpami.2022.3218569 2022
[76]

In: 2022 IEEE 24th International Workshop on Multimedia Signal Processing (MMSP), pp

Gao, L., Hu, H.-M., Li, M.: A progressive domain adaptation for object detec- tion via coarse-grained foreground guidance. In: 2022 IEEE 24th International Workshop on Multimedia Signal Processing (MMSP), pp. 01–06 (2022). https: //doi.org/10.1109/mmsp55362.2022.9948712 . IEEE 39

work page doi:10.1109/mmsp55362.2022.9948712 2022
[77]

In: The Thirty-ninth Annual Conference on Neural Information Processing Systems (2025)

Liu, C., Xiang, X., Duan, Z., Li, W., Fan, Q., Gao, Y.: Don’t need retrain- ing: A mixture of detr and vision foundation models for cross-domain few-shot object detection. In: The Thirty-ninth Annual Conference on Neural Information Processing Systems (2025)

2025
[78]

Remote Sensing17(23), 3854 (2025) https://doi.org/10

Cheng, G., Yang, H., Tian, Y., Xie, M., Dang, C., Ding, Q., Feng, X.: Wmfa-at: Adaptive teacher with weighted multi-layer feature alignment for cross-domain uav object detection. Remote Sensing17(23), 3854 (2025) https://doi.org/10. 3390/rs17233854

2025
[79]

IEEE Transactions on Circuits and Systems for Video Technology (2025) https://doi.org/10.1109/ tcsvt.2025.3538770

Wang, K., Zhou, P., Hu, M., Lu, J.: Unsupervised 3d object detection domain adaptation based on pseudo-label variance regularization. IEEE Transactions on Circuits and Systems for Video Technology (2025) https://doi.org/10.1109/ tcsvt.2025.3538770

work page arXiv 2025
[82]

arXiv preprint arXiv:2512.17514 (2025)

VCR, S., Lalla, R., Dayal, A., Kulkarni, T., Lalla, A., Balasubramanian, V.N., Khan, M.H.: Foundation model priors enhance object focus in feature space for source-free object detection. arXiv preprint arXiv:2512.17514 (2025)

work page arXiv 2025

Showing first 80 references.