Pith. sign in

REVIEW 5 major objections 6 minor 48 references

Prmpt2Adpt: Prompt-Based Zero-Shot Domain Adaptation for Resource-Constrained Environments

T0 review · 5 major / 6 minor · reviewed 2026-08-15 · deepseek-v4-flash

Challenge this review Re-run · record.json Download PDF Read on arXiv ↗ Review JSON

Pith's one-line read A text prompt plus five source images adapts a lightweight detector to unseen weather, with no target-domain images.

desk verdict Useful efficiency-driven variant of PODA with a misleading zero-shot claim and unmeasured pseudo-labels; worth a major revision. read the letter →

arxiv 2506.16994 v1 pith:DHUOONBQ submitted 2025-06-20 cs.CV cs.LG

Yasir Ali Farrukh , Syed Wali , Irfan Khan , Nathaniel D. Bastian This is my paper · ORCID
classification cs.CVcs.LG
keywords zero-shotdomainadaptationobjectdetectionprompt-drivenCLIPteacher-studentresource-constrainedaerialimagerypseudo-labeling
verification ladder T0 review T1 audit T2 compute T3 formal

The pith

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

The reading

Prmpt2Adpt tackles a practical deployment problem: a drone object detector trained in clear weather must keep working when it starts snowing, raining, fogging, or dusting, even though no images from those conditions were available at training time. The paper claims this zero-shot adaptation can be driven by a natural-language prompt alone, with no target-domain images and only five cached source images. A distilled CLIP image encoder, frozen inside a Faster R-CNN teacher, supplies semantic embeddings; a small optimization steers the statistics of source features toward the text prompt, the teacher's detection head is briefly fine-tuned, and the teacher then labels early target frames so a compact YOLOv11 nano student can adapt on the fly. On the MDS-A aerial benchmark, the method lands within a few points of heavier prompt-based baselines on snow, dust, leaves, rain, and fog while adapting up to 7x faster and inferring up to 5x faster.

What carries the argument

The load-bearing object is Prompt-driven Instance Normalization (PIN), an affine transform over the channel-wise mean and standard deviation of a source feature map. Its style parameters are learned by gradient descent so that the transformed features' embedding approaches the target prompt embedding in CLIP space under a cosine-distance loss, pulling low-level visual statistics toward the target's semantics while keeping spatial structure intact. Around PIN sits a two-model loop: a Faster R-CNN teacher with a frozen distilled CLIP backbone and trainable RPN/ROI head is fine-tuned briefly on the steered features and source labels, then generates pseudo-labels on the first target frames; those pseudo-labels adapt a YOLOv11 nano student for fast on-device inference.

What would settle it

Compute the adapted teacher's mAP directly on target-domain ground truth, or compare its pseudo-labels with ground-truth boxes on the target frames. If teacher accuracy after prompt steering is no better than the source-only baseline, the pseudo-labels carry no new target information and the student's improvement must come from some other effect; if teacher accuracy is genuinely high, the pseudo-label path is confirmed.

Watch

Extended reading notes

Core claim

The central claim is that domain adaptation for object detection can be decomposed into a cheap semantic steering step and a cheap student update, so that the entire adaptation loop fits a resource-constrained drone. Concretely, Prmpt2Adpt asserts that optimizing only the channel-wise mean and standard deviation of low-level source features, so that their CLIP embedding moves toward the embedding of a target text prompt, is enough to let a detection head trained on clear weather produce useful pseudo-labels on snow, rain, fog, dust, or falling leaves. Those pseudo-labels then transfer the adaptation to YOLOv11 nano, giving 63.6 mAP on snow, 64.5 on dust, 41.2 on leaves, 62.8 on rain, and 53.0 on fog, compared with 65.7-67.3, 67.4-70.1, 44.1-45.6, 64.0-66.2, and 55.9-58.1 for the PODA and ULDA baselines. The efficiency claim is sharper than the accuracy claim: up to 7x faster adaptation and 5x faster inference with five source images instead of full-dataset access.

Load-bearing premise

The entire pipeline rests on the teacher's detection head, fine-tuned on only five prompt-steered source images with no target validation, producing pseudo-labels on the target stream that are accurate enough to train the student; the paper calls these predictions 'high-quality' but never measures their quality.

Editorial extensions

If this is right

  • A deployed drone can switch to an unseen weather condition by swapping a text prompt, without caching target video or re-training on the full source dataset.
  • Adaptation memory drops to five stored source images, which is small enough for onboard storage in edge devices.
  • The teacher-student split concentrates expensive semantic steering in the teacher and real-time detection in the student, so the speed advantage persists after adaptation.
  • If the teacher's pseudo-labels are trustworthy, the student keeps improving during the first moments of deployment in the new domain, covering conditions that change while the drone is airborne.

Reading between the lines

Editorial extensions of the paper, not claims the author makes directly.

  • The teacher's pseudo-label quality is never measured against target ground truth; a natural follow-up is to report teacher mAP on the target domain after prompt steering, since that single number would show how much of the student's gain is real signal rather than label noise.
  • Because PIN changes only feature statistics, it is most plausible for shifts that are largely stylistic, like weather and lighting; a stress test would be a structural shift such as camera altitude or sensor type, where a text prompt may not describe the mismatch.
  • The framework's ceiling depends on the captioning pipeline generating a faithful target prompt; an informative experiment is to feed vague or wrong prompts and measure how gracefully accuracy declines.
  • The steering step is head-agnostic, so the same prompt-alignment idea could be transferred to other dense tasks such as segmentation or tracking on the same edge hardware.
Share X Bluesky LinkedIn Reddit HN

Signed reviews

No signed human review yet.

Editorial analysis

A structured set of objections, weighed in public.

Desk editor's note, referee report, and a circularity audit.

Referee Report

5 major / 6 minor

Summary. Prmpt2Adpt proposes a teacher-student framework for prompt-based domain adaptation in object detection. A frozen distilled CLIP backbone in a Faster R-CNN teacher is combined with Prompt-driven Instance Normalization (PIN) to steer low-level source features toward the target text-prompt embedding; the teacher head is briefly fine-tuned on five cached source images. During deployment, the teacher generates pseudo-labels on initial target frames, which are used to adapt a YOLOv11 nano student. Experiments on the MDS-A dataset report mAP values competitive with PODA and ULDA (e.g., 63.6 mAP for snow) and claim up to 7x faster adaptation and 5x faster inference.

Significance. The problem is timely and the efficiency-oriented goal is well motivated; if the results were properly supported, the approach would be a practical contribution to low-resource aerial detection. The paper targets a public benchmark and makes a clear efficiency argument. However, the current manuscript does not provide enough evidence for its central claims: the pseudo-label quality is not measured, the student adaptation gain is not isolated, the comparisons lack statistical support and implementation detail, and the 'zero-shot' terminology conflicts with the use of target-domain frames during adaptation. These gaps prevent assessment of whether the reported mAP numbers come from the proposed adaptation mechanism. No code or detailed configuration is provided for reproducibility.

major comments (5)
  1. [Abstract and Section 3.5] The abstract and Figure 1 claim adaptation 'without requiring target-domain images,' but Section 3.5 states that the adapted teacher is deployed on 'initial target-domain instances' to generate pseudo-labels for the student. This is a direct contradiction unless zero-shot is defined to allow target data at deployment, which the manuscript does not do. Please clarify the setting and adjust the terminology to match the actual procedure, which appears to be test-time or streaming adaptation rather than zero-shot adaptation in the usual sense.
  2. [Section 3.5] The claim that teacher predictions are 'high-quality predictions' is unsupported. The student's adaptation is entirely dependent on these pseudo-labels, yet no measurement of pseudo-label accuracy is reported, even though target ground truth is available for evaluation. Add an analysis of pseudo-label precision/recall against target annotations, and report the student's detection performance before and after teacher-guided adaptation to quantify the contribution of the pseudo-label loop.
  3. [Table 1 and Section 4.3] No error bars, confidence intervals, or number of experimental runs are given for any mAP number, including the 63.6 mAP for snow. With single-run results and small gaps to baselines (e.g., 63.6 vs. 65.7 for PODA on snow), the 'competitive' claim is not statistically grounded. Report multiple seeds with mean and standard deviation for all methods.
  4. [Section 4.2] The selection of the five representative source images is not described: no criterion, algorithm, or sensitivity analysis is provided. Since PIN optimization and teacher fine-tuning operate on exactly these five images, the results may depend strongly on the chosen images. Specify the selection procedure and report results over several selections or random draws.
  5. [Section 4.3 and Figure 6] The comparison with PODA and ULDA lacks implementation detail. These methods were originally proposed for segmentation, and the paper only states that 'analogous detection heads' were implemented. Without the exact adapted architectures, training protocol, and hyperparameters, the speed and accuracy comparisons in Table 1 and Figure 6 cannot be assessed for fairness. Provide full implementation details and, if possible, use official code or a detailed adaptation protocol.
minor comments (6)
  1. [Algorithm 1] The function 'get-embedding' in line 7 is not defined; clarify how the transformed feature map is mapped to the embedding space used in Eq. (2), and whether the full CLIP image encoder is applied to the Layer-1 feature map.
  2. [Section 3.6] No evaluation of the LLaMA-generated captions is provided; report a human or automatic quality check, because the captioning pipeline is a key input to the method and the paper asserts the result is 'high-quality.'
  3. [Section 4.1] In the dataset description, 'NA T2021' should be 'NAT2021'; also, reference [47] is for nighttime aerial tracking, so clarify why it is used for weather and scenario diversity.
  4. [Figure 5] The bars in Figure 5 do not show numerical mAP values, making the exact performance drops difficult to read; add value labels.
  5. [References] References [32] and [33] are the same paper (Radford et al.), so they should be merged or renumbered.
  6. [Section 1] The text 'UA Vs' in the introduction should read 'UAVs.'

Circularity Check

0 steps flagged · score 0.0 of 10

No significant circularity: the prompt-steering optimization, teacher fine-tuning, and student adaptation are all evaluated against held-out target ground truth.

full rationale

Prmpt2Adpt's derivation chain is not circular. The PIN feature steering (Eqs. 1-2, Algorithm 1) optimizes per-image channel-wise statistics (mu, sigma) against a fixed CLIP text embedding TrgEmb; this anchor is external to the detection pipeline and is not derived from the detection targets. The teacher's detection head is fine-tuned on five source images with their original ground-truth boxes, using the steered features, and the reported Table 1 mAP values are measured on target-domain ground truth, so no fitted parameter is renamed as a prediction. The teacher-student loop in Section 3.5 does rely on an unverified assumption that the teacher's pseudo-labels are 'high-quality predictions,' and the paper does not ablate student performance before and after pseudo-label training or measure pseudo-label accuracy; however, this is an experimental/correctness gap, not a circular reduction. The abstract's phrase 'without requiring target-domain images' is semantically loose because target frames are consumed at deployment to generate pseudo-labels, but the problem formulation explicitly scopes target-image unavailability to the training phase, so this is a terminology issue rather than a self-definitional inference. The only same-group citation is the MDS-A benchmark [29], used as an external evaluation set; it is not load-bearing for the method's derivation. Therefore no step reduces, by construction or by self-citation, to its own inputs.

Assumptions & free parameters 4 free parameters · 4 assumptions · 0 invented entities

The central claim rests on several unstated or weakly validated premises: that a single text prompt adequately represents the target domain in CLIP space, that five arbitrarily selected source images represent the source domain, that the teacher's unvalidated pseudo-labels are accurate enough for student training, and that fine-tuning on other aerial datasets transfers to MDS-A.

free parameters (4)
  • Number of cached source images = 5
    Section 4.3 states that only five representative source images are used for teacher fine-tuning, but the selection criterion is not specified.
  • PIN optimization steps N
    Algorithm 1 requires N optimization steps per feature map, but no value is reported.
  • PIN learning rate lr
    Algorithm 1 uses a learning rate for updating the style statistics, but no value is reported.
  • Teacher head fine-tuning epochs
    Section 3.5 says the head is briefly fine-tuned for a limited number of epochs, but the number is not reported.
assumptions (4)
  • domain assumption The target text prompt embedding in CLIP space captures the target domain appearance distribution well enough that steering source Layer1 features toward it improves detection.
    Section 3.4 uses a cosine loss to align steered features with the target prompt embedding; if the prompt is not representative, the steering is misdirected.
  • domain assumption The five stored source images are representative of the source domain distribution.
    Section 4.3 mentions five representative source images selected and cached before deployment, but no selection procedure is given.
  • ad hoc to paper The teacher's fine-tuned detection head produces high-quality pseudo-labels on target-domain streaming data.
    Section 3.5 calls the predictions high-quality, but no teacher-only accuracy or pseudo-label precision is reported.
  • domain assumption Fine-tuning the distilled CLIP on UAVDT, AID, VDD, and NAT2021 improves image-text alignment for MDS-A scenarios without leaking test-domain information.
    Section 3.3 uses external aerial datasets and LLaMA-generated captions; transfer to MDS-A is assumed and the filtered captions are not independently validated.

how reviews work

0 comments
Cite this review

Pith. "Pith review of Prmpt2Adpt: Prompt-Based Zero-Shot Domain Adaptation for Resource-Constrained Environments." pith.science (2026). https://pith.science/paper/DHUOONBQ

@misc{pith2026250616994,
  author       = {Pith},
  title        = {Pith review of: Prmpt2Adpt: Prompt-Based Zero-Shot Domain Adaptation for Resource-Constrained Environments},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/DHUOONBQ}},
  note         = {Machine review of arXiv:2506.16994}
}
read the original abstract

Unsupervised Domain Adaptation (UDA) is a critical challenge in real-world vision systems, especially in resource-constrained environments like drones, where memory and computation are limited. Existing prompt-driven UDA methods typically rely on large vision-language models and require full access to source-domain data during adaptation, limiting their applicability. In this work, we propose Prmpt2Adpt, a lightweight and efficient zero-shot domain adaptation framework built around a teacher-student paradigm guided by prompt-based feature alignment. At the core of our method is a distilled and fine-tuned CLIP model, used as the frozen backbone of a Faster R-CNN teacher. A small set of low-level source features is aligned to the target domain semantics-specified only through a natural language prompt-via Prompt-driven Instance Normalization (PIN). These semantically steered features are used to briefly fine-tune the detection head of the teacher model. The adapted teacher then generates high-quality pseudo-labels, which guide the on-the-fly adaptation of a compact student model. Experiments on the MDS-A dataset demonstrate that Prmpt2Adpt achieves competitive detection performance compared to state-of-the-art methods, while delivering up to 7x faster adaptation and 5x faster inference speed using few source images-making it a practical and scalable solution for real-time adaptation in low-resource domains.

Figures

Figures reproduced from arXiv: 2506.16994 by the authors.

Figure 1
Figure 1. Prmpt2Adpt addresses UDA in a zero-shot setting, [PITH_FULL_IMAGE:figures/full_fig_p001_1.png] view at source ↗
Figure 2
Figure 2. Overview of the teacher-student model architectures [PITH_FULL_IMAGE:figures/full_fig_p003_2.png] view at source ↗
Figure 3
Figure 3. Overview of our proposed Prmpt2Adpt framework. The left panel illustrates the prompt-based feature alignment, where embed [PITH_FULL_IMAGE:figures/full_fig_p004_3.png] view at source ↗
Figures from the paper (2 more)
Figure 4
Figure 4. Figure 4: Visualization of embedding distances between image [PITH_FULL_IMAGE:figures/full_fig_p005_4.png]
Figure 6
Figure 6. Figure 6: Comparison of inference and adaptation time between [PITH_FULL_IMAGE:figures/full_fig_p008_6.png]

Discussion (0). Continue with ORCID to comment.

Reference graph

Works this paper leans on

48 extracted references · 33 canonical work pages

  1. [1]

    Overcoming distribution shift in machine learning: A survey on regularization techniques for transfer learning

    Gideon Areo. Overcoming distribution shift in machine learning: A survey on regularization techniques for transfer learning. 2025. 1

    work page 2025

  2. [2]

    Vdd: Varied drone dataset for semantic segmentation.arXiv preprint arXiv:2305.13608, 2023

    Wenxiao Cai, Ke Jin, Jinyan Hou, Cong Guo, Letian Wu, and Wankou Yang. Vdd: Varied drone dataset for semantic segmentation.arXiv preprint arXiv:2305.13608, 2023. 5, 7

    arXiv 2023

  3. [3]

    Train- ing on the fly: On-device self-supervised learning aboard nano-drones within 20mw.IEEE Transactions on Computer- Aided Design of Integrated Circuits and Systems, 2024

    Elia Cereda, Alessandro Giusti, and Daniele Palossi. Train- ing on the fly: On-device self-supervised learning aboard nano-drones within 20mw.IEEE Transactions on Computer- Aided Design of Integrated Circuits and Systems, 2024. 1

    work page 2024

  4. [4]

    Harmonizing transferability and discrim- inability for adapting object detectors

    Chaoqi Chen, Zebiao Zheng, Xinghao Ding, Yue Huang, and Qi Dou. Harmonizing transferability and discrim- inability for adapting object detectors. InProceedings of the IEEE/CVF conference on computer vision and pattern recognition, pages 8869–8878, 2020. 2

    work page 2020

  5. [5]

    Flexit: Towards flexible se- mantic image translation

    Guillaume Couairon, Asya Grechka, Jakob Verbeek, Holger Schwenk, and Matthieu Cord. Flexit: Towards flexible se- mantic image translation. InProceedings of the IEEE/CVF conference on computer vision and pattern recognition, pages 18270–18279, 2022. 2

    work page 2022

  6. [6]

    Un- biased mean teacher for cross-domain object detection

    Jinhong Deng, Wen Li, Yuhua Chen, and Lixin Duan. Un- biased mean teacher for cross-domain object detection. In Proceedings of the IEEE/CVF conference on computer vi- sion and pattern recognition, pages 4091–4101, 2021. 2

    2021

  7. [7]

    Uavdt dataset, 2018

    Dawei Du, Yuankai Qi, Hongyang Yu, Yifan Yang, Kaiwen Duan, Guorong Li, Weigang Zhang, Qingming Huang, and Qi Tian. Uavdt dataset, 2018. 5, 7

    work page 2018

  8. [8]

    Boosting object detection with zero-shot day-night domain adaptation

    Zhipeng Du, Miaojing Shi, and Jiankang Deng. Boosting object detection with zero-shot day-night domain adaptation. InProceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition, pages 12666–12676, 2024. 2

    work page 2024

Show all 48 references
  1. [9]

    Taming transformers for high-resolution image synthesis

    Patrick Esser, Robin Rombach, and Bjorn Ommer. Taming transformers for high-resolution image synthesis. InPro- ceedings of the IEEE/CVF conference on computer vision and pattern recognition, pages 12873–12883, 2021. 2

    2021

  2. [10]

    Poda: Prompt-driven zero- shot domain adaptation

    Mohammad Fahes, Tuan-Hung Vu, Andrei Bursuc, Patrick P´erez, and Raoul De Charette. Poda: Prompt-driven zero- shot domain adaptation. InProceedings of the IEEE/CVF International Conference on Computer Vision, pages 18623– 18633, 2023. 3, 5, 8

    2023

  3. [11]

    Stylegan-nada: Clip- guided domain adaptation of image generators.ACM Trans- actions on Graphics (TOG), 41(4):1–13, 2022

    Rinon Gal, Or Patashnik, Haggai Maron, Amit H Bermano, Gal Chechik, and Daniel Cohen-Or. Stylegan-nada: Clip- guided domain adaptation of image generators.ACM Trans- actions on Graphics (TOG), 41(4):1–13, 2022. 2

    2022

  4. [12]

    Domain-adversarial training of neural networks.Journal of machine learning research, 17(59):1–35, 2016

    Yaroslav Ganin, Evgeniya Ustinova, Hana Ajakan, Pas- cal Germain, Hugo Larochelle, Franc ¸ois Laviolette, Mario March, and Victor Lempitsky. Domain-adversarial training of neural networks.Journal of machine learning research, 17(59):1–35, 2016. 2

    2016

  5. [13]

    Domain adaptation via prompt learning.IEEE Transactions on Neural Networks and Learning Systems, 2023

    Chunjiang Ge, Rui Huang, Mixue Xie, Zihang Lai, Shiji Song, Shuang Li, and Gao Huang. Domain adaptation via prompt learning.IEEE Transactions on Neural Networks and Learning Systems, 2023. 2

    2023

  6. [14]

    Cycada: Cycle-consistent adversarial domain adaptation

    Judy Hoffman, Eric Tzeng, Taesung Park, Jun-Yan Zhu, Phillip Isola, Kate Saenko, Alexei Efros, and Trevor Darrell. Cycada: Cycle-consistent adversarial domain adaptation. In International conference on machine learning, pages 1989–

    1989

  7. [15]

    Arbitrary style transfer in real-time with adaptive instance normalization

    Xun Huang and Serge Belongie. Arbitrary style transfer in real-time with adaptive instance normalization. InProceed- ings of the IEEE international conference on computer vi- sion, pages 1501–1510, 2017. 5

    2017

  8. [16]

    Scaling up visual and vision-language representa- tion learning with noisy text supervision

    Chao Jia, Yinfei Yang, Ye Xia, Yi-Ting Chen, Zarana Parekh, Hieu Pham, Quoc Le, Yun-Hsuan Sung, Zhen Li, and Tom Duerig. Scaling up visual and vision-language representa- tion learning with noisy text supervision. InInternational conference on machine learning, pages 4904–4916. PMLR,

  9. [17]

    Ultralytics yolo11, 2024

    Glenn Jocher and Jing Qiu. Ultralytics yolo11, 2024. 4, 5

    2024

  10. [18]

    A style-based generator architecture for generative adversarial networks

    Tero Karras, Samuli Laine, and Timo Aila. A style-based generator architecture for generative adversarial networks. InProceedings of the IEEE/CVF conference on computer vi- sion and pattern recognition, pages 4401–4410, 2019. 2

    2019

  11. [19]

    A review of domain adap- tation without target labels.IEEE transactions on pattern analysis and machine intelligence, 43(3):766–785, 2019

    Wouter M Kouw and Marco Loog. A review of domain adap- tation without target labels.IEEE transactions on pattern analysis and machine intelligence, 43(3):766–785, 2019. 1

    2019

  12. [20]

    Clipstyler: Image style transfer with a single text condition

    Gihyun Kwon and Jong Chul Ye. Clipstyler: Image style transfer with a single text condition. InProceedings of the IEEE/CVF conference on computer vision and pattern recognition, pages 18062–18071, 2022. 2

    2022

  13. [21]

    Zero-shot day-night domain adaptation with a physics prior

    Attila Lengyel, Sourav Garg, Michael Milford, and Jan C van Gemert. Zero-shot day-night domain adaptation with a physics prior. InProceedings of the IEEE/CVF International Conference on Computer Vision, pages 4399–4409, 2021. 2

    2021

  14. [22]

    Align before fuse: Vision and language representation learn- ing with momentum distillation.Advances in neural infor- mation processing systems, 34:9694–9705, 2021

    Junnan Li, Ramprasaath Selvaraju, Akhilesh Gotmare, Shafiq Joty, Caiming Xiong, and Steven Chu Hong Hoi. Align before fuse: Vision and language representation learn- ing with momentum distillation.Advances in neural infor- mation processing systems, 34:9694–9705, 2021. 2

    2021

  15. [23]

    Bidirectional learning for domain adaptation of semantic segmentation

    Yunsheng Li, Lu Yuan, and Nuno Vasconcelos. Bidirectional learning for domain adaptation of semantic segmentation. In Proceedings of the IEEE/CVF conference on computer vi- sion and pattern recognition, pages 6936–6945, 2019. 2

    2019

  16. [24]

    Cross-domain adaptive teacher for object detection

    Yu-Jhe Li, Xiaoliang Dai, Chih-Yao Ma, Yen-Cheng Liu, Kan Chen, Bichen Wu, Zijian He, Kris Kitani, and Peter Vajda. Cross-domain adaptive teacher for object detection. InProceedings of the IEEE/CVF conference on computer vi- sion and pattern recognition, pages 7581–7590, 2022. 2

    2022

  17. [25]

    Conditional adversarial domain adapta- tion.Advances in neural information processing systems, 31, 2018

    Mingsheng Long, Zhangjie Cao, Jianmin Wang, and Michael I Jordan. Conditional adversarial domain adapta- tion.Advances in neural information processing systems, 31, 2018. 2

    2018

  18. [26]

    Exploring models and data for remote sensing im- age caption generation.IEEE Transactions on Geoscience and Remote Sensing, 56(4):2183–2195

    Xiaoqiang Lu, Binqiang Wang, Xiangtao Zheng, and Xue- long Li. Exploring models and data for remote sensing im- age caption generation.IEEE Transactions on Geoscience and Remote Sensing, 56(4):2183–2195. 6

  19. [27]

    Adversarial style mining for one-shot unsupervised domain adaptation.Advances in neural information processing sys- tems, 33:20612–20623, 2020

    Yawei Luo, Ping Liu, Tao Guan, Junqing Yu, and Yi Yang. Adversarial style mining for one-shot unsupervised domain adaptation.Advances in neural information processing sys- tems, 33:20612–20623, 2020. 2

    2020

  20. [28]

    Llama 3.2: Revolutionizing edge ai and vision with open, customizable models.Meta AI, 2024

    AI Meta. Llama 3.2: Revolutionizing edge ai and vision with open, customizable models.Meta AI, 2024. 6

    2024

  21. [29]

    Multiple distribution shift–aerial (mds- a): A dataset for test-time error detection and model adapta- tion.arXiv preprint arXiv:2502.13289, 2025

    Noel Ngu, Aditya Taparia, Gerardo I Simari, Mario Leiva, Jack Corcoran, Ransalu Senanayake, Paulo Shakarian, and Nathaniel D Bastian. Multiple distribution shift–aerial (mds- a): A dataset for test-time error detection and model adapta- tion.arXiv preprint arXiv:2502.13289, 2025. 7

    2025 arXiv

  22. [30]

    Unsupervised intra-domain adaptation for se- mantic segmentation through self-supervision

    Fei Pan, Inkyu Shin, Francois Rameau, Seokju Lee, and In So Kweon. Unsupervised intra-domain adaptation for se- mantic segmentation through self-supervision. InProceed- ings of the IEEE/CVF conference on computer vision and pattern recognition, pages 3764–3773, 2020. 2

    2020

  23. [31]

    Styleclip: Text-driven manipulation of stylegan imagery

    Or Patashnik, Zongze Wu, Eli Shechtman, Daniel Cohen-Or, and Dani Lischinski. Styleclip: Text-driven manipulation of stylegan imagery. InProceedings of the IEEE/CVF inter- national conference on computer vision, pages 2085–2094,

  24. [33]

    Learning transferable visual models from natural language supervi- sion

    Alec Radford, Jong Wook Kim, Chris Hallacy, Aditya Ramesh, Gabriel Goh, Sandhini Agarwal, Girish Sastry, Amanda Askell, Pamela Mishkin, Jack Clark, et al. Learning transferable visual models from natural language supervi- sion. InInternational conference on machine learning, p...

    2021

  25. [34]

    Strong-weak distribution alignment for adaptive ob- ject detection

    Kuniaki Saito, Yoshitaka Ushiku, Tatsuya Harada, and Kate Saenko. Strong-weak distribution alignment for adaptive ob- ject detection. InProceedings of the IEEE/CVF conference on computer vision and pattern recognition, pages 6956– 6965, 2019. 2

    2019

  26. [35]

    Scl: Towards accurate domain adaptive object detection via gradient detach based stacked complementary losses.arXiv preprint arXiv:1911.02559, 2019

    Zhiqiang Shen, Harsh Maheshwari, Weichen Yao, and Mar- ios Savvides. Scl: Towards accurate domain adaptive object detection via gradient detach based stacked complementary losses.arXiv preprint arXiv:1911.02559, 2019. 2

    1911 arXiv

  27. [36]

    Learning to adapt structured output space for semantic seg- mentation

    Yi-Hsuan Tsai, Wei-Chih Hung, Samuel Schulter, Ki- hyuk Sohn, Ming-Hsuan Yang, and Manmohan Chandraker. Learning to adapt structured output space for semantic seg- mentation. InProceedings of the IEEE conference on computer vision and pattern recognition, pages 7472–7481,

  28. [37]

    Measuring domain shifts using deep learning remote photoplethysmography model similarity.arXiv preprint arXiv:2404.08184, 2024

    Nathan Vance and Patrick Flynn. Measuring domain shifts using deep learning remote photoplethysmography model similarity.arXiv preprint arXiv:2404.08184, 2024. 1

    2024 arXiv

  29. [38]

    Mega-cda: Memory guided attention for category-aware unsupervised domain adaptive object detection

    Vibashan Vs, Vikram Gupta, Poojan Oza, Vishwanath A Sindagi, and Vishal M Patel. Mega-cda: Memory guided attention for category-aware unsupervised domain adaptive object detection. InProceedings of the IEEE/CVF Confer- ence on Computer Vision and Pattern Recognition, pages 451...

    2021

  30. [39]

    Advent: Adversarial entropy min- imization for domain adaptation in semantic segmentation

    Tuan-Hung Vu, Himalaya Jain, Maxime Bucher, Matthieu Cord, and Patrick P ´erez. Advent: Adversarial entropy min- imization for domain adaptation in semantic segmentation. InProceedings of the IEEE/CVF conference on computer vi- sion and pattern recognition, pages 2517–2526, 2019. 2

    2019

  31. [40]

    One-shot unsupervised domain adaptation for object detec- tion

    Zhiqiang Wan, Lusi Li, Hepeng Li, Haibo He, and Zhen Ni. One-shot unsupervised domain adaptation for object detec- tion. In2020 International Joint Conference on Neural Net- works (IJCNN), pages 1–8. IEEE, 2020. 2

    2020

  32. [41]

    Deep domain adaptation by geodesic distance minimization

    Yifei Wang, Wen Li, Dengxin Dai, and Luc Van Gool. Deep domain adaptation by geodesic distance minimization. In Proceedings of the IEEE international conference on com- puter vision workshops, pages 2651–2657, 2017. 2

    2017

  33. [42]

    Tinyclip: Clip dis- tillation via affinity mimicking and weight inheritance

    Kan Wu, Houwen Peng, Zhenghong Zhou, Bin Xiao, Mengchen Liu, Lu Yuan, Hong Xuan, Michael Valenzuela, Xi Stephen Chen, Xinggang Wang, et al. Tinyclip: Clip dis- tillation via affinity mimicking and weight inheritance. In Proceedings of the IEEE/CVF International Conference on C...

    2023

  34. [43]

    Style mixing and patchwise prototypical matching for one- shot unsupervised domain adaptive semantic segmentation

    Xinyi Wu, Zhenyao Wu, Yuhang Lu, Lili Ju, and Song Wang. Style mixing and patchwise prototypical matching for one- shot unsupervised domain adaptive semantic segmentation. InProceedings of the AAAI Conference on Artificial Intelli- gence, pages 2740–2749, 2022. 2

    2022

  35. [44]

    Aid: A benchmark data set for performance evaluation of aerial scene classification.IEEE Transactions on Geoscience and Remote Sensing, 55(7):3965–3981, 2017

    Gui-Song Xia, Jingwen Hu, Fan Hu, Baoguang Shi, Xiang Bai, Yanfei Zhong, Liangpei Zhang, and Xiaoqiang Lu. Aid: A benchmark data set for performance evaluation of aerial scene classification.IEEE Transactions on Geoscience and Remote Sensing, 55(7):3965–3981, 2017. 5, 7

    2017

  36. [45]

    Unified language-driven zero-shot domain adaptation

    Senqiao Yang, Zhuotao Tian, Li Jiang, and Jiaya Jia. Unified language-driven zero-shot domain adaptation. InProceed- ings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition, pages 23407–23415, 2024. 3, 8

    2024

  37. [46]

    Fda: Fourier domain adaptation for semantic segmentation

    Yanchao Yang and Stefano Soatto. Fda: Fourier domain adaptation for semantic segmentation. InProceedings of the IEEE/CVF conference on computer vision and pattern recognition, pages 4085–4095, 2020. 2

    2020

  38. [47]

    Unsupervised domain adap- tation for nighttime aerial tracking

    Junjie Ye, Changhong Fu, Guangze Zheng, Danda Pani Paudel, and Guang Chen. Unsupervised domain adap- tation for nighttime aerial tracking. InProceedings of the IEEE/CVF conference on computer vision and pattern recognition, pages 8896–8905, 2022. 5, 7

    2022

  39. [48]

    Adapting object detectors via selective cross- domain alignment

    Xinge Zhu, Jiangmiao Pang, Ceyuan Yang, Jianping Shi, and Dahua Lin. Adapting object detectors via selective cross- domain alignment. InProceedings of the IEEE/CVF con- ference on computer vision and pattern recognition, pages 687–696, 2019. 2

    2019

  40. [49]

    Confidence regularized self-training

    Yang Zou, Zhiding Yu, Xiaofeng Liu, BVK Kumar, and Jin- song Wang. Confidence regularized self-training. InPro- ceedings of the IEEE/CVF international conference on com- puter vision, pages 5982–5991, 2019. 2

    2019

Pith tools

Reviewed August 15, 2026 · model on record in the stance chip above.