Pith. sign in

REVIEW 4 major objections 5 minor 50 references

SenCLIP: Enhancing zero-shot land-use mapping for Sentinel-2 with ground-level prompting

T0 review · 4 major / 5 minor · reviewed 2026-08-11 · deepseek-v4-flash

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

Pith's one-line read SenCLIP aligns Sentinel-2 imagery with co-located ground-level photos and claims consistent zero-shot LULC gains over remote sensing VLMs on EuroSAT and BigEarthNet

desk verdict Useful extension of Sat2Cap to Sentinel-2, but the main comparison is undermined by an under-specified transductive normalization step; the core idea is still worth engaging with. read the letter →

arxiv 2412.08536 v1 pith:4NF2FNUH submitted 2024-12-11 cs.CV

Pallavi Jain , Dino Ienco , Roberto Interdonato , Tristan Berchoux , Diego Marcos This is my paper · ORCID
classification cs.CV
keywords zero-shotland-use/land-coverclassificationcross-viewrepresentationlearningSentinel-2CLIPfine-tuningground-levelpromptingcontrastiveLUCASdatasetremotesensingvision-languagemodels
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

SenCLIP aims to make CLIP useful for zero-shot land-use and land-cover (LULC) mapping from Sentinel-2's 10 m imagery, a domain where CLIP's ground-level pretraining is a poor match. The paper's idea is to fine-tune CLIP's satellite image encoder so that Sentinel-2 patches are pulled toward the CLIP embeddings of co-located, geotagged ground-level photos from the LUCAS survey. Because those ground photos are already well aligned with CLIP's text space, a satellite patch inherits the ability to be compared with free-form textual descriptions such as 'a field with rows of young plants' alongside standard 'aerial photo of' phrases. The paper reports consistent accuracy gains over CLIP and remote-sensing VLMs on EuroSAT and BigEarthNet under both aerial and ground-level prompts, and introduces an LLM-based prompt-selection step that concentrates class descriptions into a few effective prompts. If the results hold, zero-shot LULC mapping becomes practical without labeled training data or curated remote-sensing captions.

What carries the argument

The central mechanism is a cross-view contrastive alignment: frozen ground-level CLIP embeddings $G_i$ (average- or attention-pooled over the four directional LUCAS photos) act as targets for a fine-tuned Sentinel-2 encoder $f_S$ with projection head $H$, optimized by the InfoNCE loss $$\mathcal{L}_{\text{InfoNCE}} = -\frac{1}{N}\sum_{i=1}^N \log \frac{\exp(G_i \cdot H(f_S(x_i))/\tau)}{\sum_{j=1}^N \exp(G_i \cdot H(f_S(x_j))/\tau)}.$$ This trains the satellite branch to occupy the same semantic manifold as ground photos, which are already text-aligned in CLIP. The second component is the zero-shot decision rule: LLM-generated view-specific prompts are treated as class attributes, and Direct Attribute Prediction scores each class by the product of per-prompt similarities normalized by each prompt's mean similarity to the full image set. A prompt-selection step ranks prompts by the ratio of within-class to overall text similarity, so a small set of discriminative prompts can replace the full set.

What would settle it

Re-run the zero-shot evaluation with the prompt-calibration statistics $p(a_{c,t})$ computed only from a held-out set of images that is never used for evaluation, for SenCLIP and all baselines. If SenCLIP's margin over CLIP, RemoteCLIP, or SkyCLIP shrinks or reverses under this inductive protocol, the headline improvement is an artifact of transductive normalization rather than the cross-view alignment.

Watch

Extended reading notes

Core claim

The paper claims that cross-view alignment alone—without labels or captions—can transfer CLIP's ground-level semantics to medium-resolution satellite imagery. Concretely, SenCLIP freezes a CLIP image encoder for ground photos, pools the four directional LUCAS images at each location, and trains a satellite encoder plus projection head to match those pooled embeddings through an InfoNCE contrastive loss with a queue-based contrastive dictionary. At inference, class scores come from Direct Attribute Prediction over 50 LLM-generated prompts per class per view, with a prompt-selection ratio that ranks prompts by within-class versus overall similarity. The reported result is that SenCLIP outperforms CLIP, RemoteCLIP, SkyCLIP, and even the supervised GeoRSCLIP on EuroSAT accuracy and BigEarthNet mean average precision, with the largest margins on ground-level prompts. The paper positions this as evidence that ground-level prompting, not just generic satellite phrases, can drive zero-shot LULC classification.

Load-bearing premise

The comparison assumes that using the test images themselves to calibrate prompt scores helps every model equally; if SenCLIP happens to benefit more from that calibration than the baselines, the reported gains are not actually due to the ground-to-satellite alignment.

Editorial extensions

If this is right

  • Zero-shot LULC classification on EuroSAT and BigEarthNet improves under generic, aerial, and ground-level prompts compared with the tested baselines.
  • Ground-level prompts, which are usually weak for satellite CLIP models, become competitive with or better than aerial prompts for SenCLIP.
  • A small set of 2–5 selected prompts per class performs as well as or better than the full 50-prompt set on EuroSAT, so prompt curation can reduce inference cost.
  • The learned representation supports satellite-to-ground and ground-to-satellite retrieval, retrieving LUCAS ground photos that match EuroSAT classes.
  • Label-free classifier tuning converges in fewer epochs on SenCLIP than on CLIP, suggesting the representation is already aligned with the task.

Reading between the lines

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

  • Not tested in the paper: an inductive variant that estimates prompt-calibration statistics without the test set could turn SenCLIP into a strictly zero-shot rather than transductive classifier; comparing that variant against the reported numbers would separate the alignment gain from the normalization gain.
  • The same cross-view objective could be applied to urban street-view imagery paired with Sentinel-2 to build zero-shot urban LULC maps, since the method only needs co-located ground photos, not captions.
  • The prompt-selection ratio is a general text-side curation rule that could be applied to any VLM's prompt ensemble; the paper only demonstrates it on SenCLIP and EuroSAT.
  • Because BigEarthNet is multi-label and the Direct Attribute Prediction rule as written uses an argmax over classes, a per-class thresholding variant might change the reported mAP ranking; the paper does not specify this adaptation.
Share X Bluesky LinkedIn Reddit HN

Editorial analysis

A structured set of objections, weighed in public.

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

Referee Report

4 major / 5 minor

Summary. The paper proposes SenCLIP, a method that fine-tunes a CLIP image encoder on Sentinel-2 imagery paired with co-located ground-level photographs from the LUCAS dataset, using an InfoNCE contrastive objective to align satellite representations with frozen CLIP embeddings of ground-level images. The authors evaluate zero-shot land-use/land-cover classification on EuroSAT and BigEarthNet under three prompt styles (generic, aerial, ground) and also introduce a text-based prompt selection method. They claim that SenCLIP consistently outperforms existing remote sensing VLMs such as CLIP, RemoteCLIP, SkyCLIP, and GeoRSCLIP, and that the cross-view alignment enables large gains with ground-level prompts.

Significance. If the empirical claims are substantiated, the cross-view alignment idea is valuable: it transfers ground-level semantic knowledge to medium-resolution Sentinel-2 imagery without using benchmark labels, which is relevant for real-world zero-shot LULC mapping. The release of code, dataset, and pretrained models is a positive step for reproducibility. The prompt selection method is simple and appears label-free. However, the current evaluation protocol has load-bearing weaknesses that prevent the paper from supporting its central claim as written: the transductive DAP normalization in Eq. (5) is not shown to be applied identically to baselines, the multi-label inference for BigEarthNet is unspecified, and the main table lacks error bars. These issues need to be addressed before the claimed improvements can be attributed to the proposed alignment.

major comments (4)
  1. [§3.3, Eq. (5)] The DAP inference rule in Eq. (5) uses p(a_{c,t}) 'empirically estimated as the mean similarity of a_{c,t} with the full image set.' In the evaluation context this is the test set, making the procedure transductive: each test image's score depends on statistics of the entire test collection. The paper does not state whether this normalization was applied identically to CLIP, RemoteCLIP, SkyCLIP, and GeoRSCLIP in Table 1. If it was not, the comparison is not apples-to-apples; if it was, the normalization is model-dependent and can change relative rankings. Please specify the exact inference protocol used for every model and justify why the transductive estimate does not advantage SenCLIP.
  2. [Table 1, §4.2] The claim that 'SenCLIP consistently outperforms all other models' is directly contradicted by Table 1. On EuroSAT with ViT-B/32 and generic prompts, GeoRSCLIP achieves 63.40 while SenCLIP-AvgPool and SenCLIP-AttPool achieve 61.18 and 62.24, respectively. On BigEarthNet with ViT-B/32 and generic prompts, GeoRSCLIP* reaches 41.95 versus SenCLIP-AvgPool 34.72 and SenCLIP-AttPool 33.78. Even if GeoRSCLIP is excluded as 'supervised' on BigEarthNet, the EuroSAT generic-prompt results are still a counterexample. Please revise the claim to be consistent with the reported numbers, or restrict it to the prompt styles and baselines where the data actually support it.
  3. [§4.2, BigEarthNet evaluation] The paper reports mAP for BigEarthNet, which is a multi-label dataset with 19 classes, but the only inference rule given is the argmax in Eq. (5), which is appropriate for single-label classification. No per-class scoring rule or threshold is described for multi-label prediction. Without a fully specified inference protocol for BigEarthNet, the reported mAP numbers are not reproducible and the comparison to baselines is not well-defined. Please provide the exact multi-label inference procedure.
  4. [Table 1, overall] Table 1 reports single-run accuracies with no error bars, even though margins between SenCLIP and competitive baselines are small in several settings (e.g., EuroSAT ViT ground: SenCLIP-AvgPool 65.54 vs. GeoRSCLIP 65.82; SenCLIP-AttPool 66.91). The paper's 'consistently outperforms' claim needs statistical support. Please report the mean and standard deviation over multiple runs, or at least include significance tests for the main comparisons, especially where the gap is below a few points.
minor comments (5)
  1. [Abstract and §4.2] The abstract and Section 4.2 use the phrase 'significant improvements' without any statistical significance testing; please temper the wording or add appropriate tests.
  2. [§3.3, Eqs. (6)-(7)] The definitions of α_{c,t} and β_{c,t} should clarify that the dot products are between normalized text embeddings; otherwise the notation is ambiguous. Also check the denominator in Eq. (7): it currently reads C·T but the sum is over d and q, which should correspond to C*T terms.
  3. [Fig. 2] The terms 'Best K' and 'Worst K' are used but not formally defined in the text; please explain how the ordering is computed and how 'worst' prompts are selected.
  4. [§3.2] The projection head and attention pooling details (e.g., output dimensions, whether the attention weights are normalized across the four directions) are not fully specified; please add these implementation details for reproducibility.
  5. [Table 2] The LaFTer experiments report standard deviations but do not include RemoteCLIP or SkyCLIP, making it unclear whether the advantage over these baselines persists when the text classifier is tuned; please clarify the scope of this comparison.

Circularity Check

0 steps flagged · score 0.0 of 10

No circular derivation: SenCLIP's label-free cross-view training is independent of the EuroSAT/BigEarthNet labels; the only test-set dependence is a transductive DAP normalizer, which is a fairness caveat rather than a circular step.

full rationale

The derivation chain is not circular. SenCLIP is trained with an InfoNCE contrastive loss (Eq. 4) that aligns a fine-tuned Sentinel-2 image encoder with frozen CLIP embeddings of co-located LUCAS ground-level photos; neither EuroSAT nor BigEarthNet labels or captions enter this objective, so the representation learning is self-contained and label-free. Prompt generation and the prompt selection scores (Eqs. 6-7) are computed from text embeddings only, not from benchmark images or labels. The one step that touches the evaluation set is the DAP normalization in Eq. (5), where p(ac,t) is 'empirically estimated as the mean similarity of ac,t with the full image set'; this makes inference transductive and should be controlled for across baselines, but it is a test-set statistic rather than a fitted label predictor and does not make the accuracy claim equivalent to the training input. The comparison also has non-circular reporting gaps: BigEarthNet is multi-label yet Eq. (5) is an argmax rule while mAP is reported, and GeoRSCLIP is excluded from 'consistent outperformance' only by a supervision caveat. No load-bearing self-citation or imported uniqueness claim appears; references to Sat2Cap, DAP, MoCo, and LaFTer are independent prior work. Accordingly, no circular step is recorded.

Assumptions & free parameters 2 free parameters · 3 assumptions · 0 invented entities

The method relies on standard CLIP components and the LUCAS geotagged photo collection. No new physical entities are introduced. The main assumptions are domain-level, about the quality of CLIP's ground-level semantic space and the representativeness of LUCAS.

free parameters (2)
  • temperature tau = 0.07
    Scaling temperature in InfoNCE loss, set by hand following CLIP convention; not fitted to the target benchmark but affects training.
  • number of prompts per class T = 50
    Chosen by the authors for all classes and both views; the prompt selection method then selects a subset, and Figure 2 explores different K values.
assumptions (3)
  • domain assumption CLIP's frozen ground-level image embeddings provide a semantic space in which text prompts for LULC classes align with visual features.
    The entire method relies on the pretrained CLIP text-image alignment for ground-level photos, which is the target space that satellite embeddings are trained to match.
  • domain assumption The LUCAS 2018 ground-level photos are representative of the land-use and land-cover classes in EuroSAT and BigEarthNet.
    The cross-view training uses LUCAS locations; if the visual appearance or class distribution of LUCAS differs from the benchmarks, the transferred representation may not generalize.
  • standard math DAP (Lampert et al. 2009) is a valid way to combine per-attribute scores into class predictions.
    The paper uses the DAP rule without modification or justification for the zero-shot multi-label setting.

how reviews work

0 comments
Cite this review

Pith. "Pith review of SenCLIP: Enhancing zero-shot land-use mapping for Sentinel-2 with ground-level prompting." pith.science (2026). https://pith.science/paper/4NF2FNUH

@misc{pith2026241208536,
  author       = {Pith},
  title        = {Pith review of: SenCLIP: Enhancing zero-shot land-use mapping for Sentinel-2 with ground-level prompting},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/4NF2FNUH}},
  note         = {Machine review of arXiv:2412.08536}
}
read the original abstract

Pre-trained vision-language models (VLMs), such as CLIP, demonstrate impressive zero-shot classification capabilities with free-form prompts and even show some generalization in specialized domains. However, their performance on satellite imagery is limited due to the underrepresentation of such data in their training sets, which predominantly consist of ground-level images. Existing prompting techniques for satellite imagery are often restricted to generic phrases like a satellite image of ..., limiting their effectiveness for zero-shot land-use and land-cover (LULC) mapping. To address these challenges, we introduce SenCLIP, which transfers CLIPs representation to Sentinel-2 imagery by leveraging a large dataset of Sentinel-2 images paired with geotagged ground-level photos from across Europe. We evaluate SenCLIP alongside other SOTA remote sensing VLMs on zero-shot LULC mapping tasks using the EuroSAT and BigEarthNet datasets with both aerial and ground-level prompting styles. Our approach, which aligns ground-level representations with satellite imagery, demonstrates significant improvements in classification accuracy across both prompt styles, opening new possibilities for applying free-form textual descriptions in zero-shot LULC mapping.

Figures

Figures reproduced from arXiv: 2412.08536 by the authors.

Figure 1
Figure 1. Architecture: The figure illustrates the three-step pipeline consisting of Pre-Training, Prompt Selection, and Zero-shot Predic [PITH_FULL_IMAGE:figures/full_fig_p004_1.png] view at source ↗
Figure 2
Figure 2. Effect of prompt selection strategies on model (RN50 [PITH_FULL_IMAGE:figures/full_fig_p006_2.png] view at source ↗
Figure 3
Figure 3. Image captioning on EuroSAT images using ClipCap [ [PITH_FULL_IMAGE:figures/full_fig_p007_3.png] view at source ↗
Figures from the paper (1 more)
Figure 4
Figure 4. Figure 4: Qualitative image-to-image retrieval: This analysis demonstrates the qualitative effectiveness of SenCLIP embeddings in both directions. By identifying the top-2 nearest LUCAS embeddings from EuroSAT images, the results indicate that the model successfully learns the f…

Discussion (0). Continue with ORCID to comment.

Reference graph

Works this paper leans on

50 extracted references · 30 canonical work pages

  1. [1]

    A simple zero-shot prompt weight- ing technique to improve prompt ensembling in text-image models

    James Urquhart Allingham, Jie Ren, Michael W Dusenberry, Xiuye Gu, Yin Cui, Dustin Tran, Jeremiah Zhe Liu, and Bal- aji Lakshminarayanan. A simple zero-shot prompt weight- ing technique to improve prompt ensembling in text-image models. In International Conference on Machine Learning, pages 547–568. PMLR, 2023. 2

    work page 2023

  2. [2]

    A land use and land cover classi- fication system for use with remote sensor data, volume 964

    James Richard Anderson. A land use and land cover classi- fication system for use with remote sensor data, volume 964. US Government Printing Office, 1976. 1

    work page 1976

  3. [3]

    Corine land cover and land cover change products

    Gy ¨orgy B ¨uttner. Corine land cover and land cover change products. In Land use and land cover mapping in Europe: practices & trends, pages 55–74. Springer, 2014. 2

    work page 2014

  4. [4]

    Sat2Cap: Mapping Fine-Grained Textual Descriptions from Satellite Images

    Aayush Dhakal, Adeel Ahmad, Subash Khanal, Srikumar Sastry, and Nathan Jacobs. Sat2cap: Mapping fine-grained textual descriptions from satellite images. arXiv preprint arXiv:2307.15904, 2023. 2, 3

    work page Pith review arXiv 2023

  5. [5]

    Lanczos filtering in one and two dimen- sions

    Claude E Duchon. Lanczos filtering in one and two dimen- sions. Journal of Applied Meteorology and Climatology , 18(8):1016–1022, 1979. 11

    work page 1979

  6. [6]

    Harmonised lucas in-situ land cover and use database for field surveys from 2006 to 2018 in the european union

    Rapha ¨el d’Andrimont, Momchil Yordanov, Laura Martinez- Sanchez, Beatrice Eiselt, Alessandra Palmieri, Paolo Do- minici, Javier Gallego, Hannes Isaak Reuter, Christian Joe- bges, Guido Lemoine, et al. Harmonised lucas in-situ land cover and use database for field surveys from 2006 to 2018 in the european union. Scientific data, 7(1):352, 2020. 3, 11

    work page 2006

  7. [7]

    Knowledge generation using satellite earth ob- servations to support sustainable development goals (sdg): A use case on land degradation

    Gregory Giuliani, Paolo Mazzetti, Mattia Santoro, Stefano Nativi, Joost Van Bemmelen, Guido Colangeli, and Anthony Lehmann. Knowledge generation using satellite earth ob- servations to support sustainable development goals (sdg): A use case on land degradation. International Journal of Applied Earth Observation and Geoinformation, 88:102068,

    work page

  8. [8]

    High-resolution global maps of 21st-century forest cover change

    Matthew C Hansen, Peter V Potapov, Rebecca Moore, Matt Hancher, Svetlana A Turubanova, Alexandra Tyukavina, David Thau, Stephen V Stehman, Scott J Goetz, Thomas R Loveland, et al. High-resolution global maps of 21st-century forest cover change. science, 342(6160):850–853, 2013. 2

    work page 2013

Show all 50 references
  1. [9]

    Hargreaves and Gary R

    Peter K. Hargreaves and Gary R. Watmough. Satellite earth observation to support sustainable rural development. Inter- national Journal of Applied Earth Observation and Geoin- formation, 103:102466, 2021. 1

    2021

  2. [10]

    Momentum contrast for unsupervised visual rep- resentation learning

    Kaiming He, Haoqi Fan, Yuxin Wu, Saining Xie, and Ross Girshick. Momentum contrast for unsupervised visual rep- resentation learning. In Proceedings of the IEEE/CVF con- ference on computer vision and pattern recognition , pages 9729–9738, 2020. 4

    2020

  3. [11]

    Deep residual learning for image recognition

    Kaiming He, Xiangyu Zhang, Shaoqing Ren, and Jian Sun. Deep residual learning for image recognition. In Proceed- ings of the IEEE conference on computer vision and pattern recognition, pages 770–778, 2016. 1

    2016

  4. [12]

    Eurosat: A novel dataset and deep learning benchmark for land use and land cover classification

    Patrick Helber, Benjamin Bischke, Andreas Dengel, and Damian Borth. Eurosat: A novel dataset and deep learning benchmark for land use and land cover classification. IEEE Journal of Selected Topics in Applied Earth Observations and Remote Sensing, 12(7):2217–2226, 2019. 5

    2019

  5. [13]

    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. In International conference on machine learning, pages 4904–4916. PMLR,

  6. [14]

    Imagenet classification with deep convolutional neural net- works

    Alex Krizhevsky, Ilya Sutskever, and Geoffrey E Hinton. Imagenet classification with deep convolutional neural net- works. Advances in neural information processing systems , 25, 2012. 1

    2012

  7. [15]

    Learning to detect unseen object classes by between- class attribute transfer

    Christoph H Lampert, Hannes Nickisch, and Stefan Harmel- ing. Learning to detect unseen object classes by between- class attribute transfer. In 2009 IEEE conference on com- puter vision and pattern recognition, pages 951–958. IEEE,

    2009

  8. [16]

    Zero-shot scene classification for high spatial reso- lution remote sensing images

    Aoxue Li, Zhiwu Lu, Liwei Wang, Tao Xiang, and Ji-Rong Wen. Zero-shot scene classification for high spatial reso- lution remote sensing images. IEEE Transactions on Geo- science and Remote Sensing, 55(7):4157–4167, 2017. 2

    2017

  9. [17]

    Blip-2: Bootstrapping language-image pre-training with frozen image encoders and large language models

    Junnan Li, Dongxu Li, Silvio Savarese, and Steven Hoi. Blip-2: Bootstrapping language-image pre-training with frozen image encoders and large language models. arXiv preprint arXiv:2301.12597, 2023. 1

    2023 arXiv

  10. [18]

    Blip: Bootstrapping language-image pre-training for uni- fied vision-language understanding and generation

    Junnan Li, Dongxu Li, Caiming Xiong, and Steven Hoi. Blip: Bootstrapping language-image pre-training for uni- fied vision-language understanding and generation. In In- ternational Conference on Machine Learning, pages 12888– 12900. PMLR, 2022. 1

    2022

  11. [19]

    Rs-clip: Zero shot remote sensing scene classification via contrastive vision-language supervision

    Xiang Li, Congcong Wen, Yuan Hu, and Nan Zhou. Rs-clip: Zero shot remote sensing scene classification via contrastive vision-language supervision. International Journal of Ap- plied Earth Observation and Geoinformation , 124:103497,

  12. [20]

    Robust deep alignment network with remote sensing knowledge graph for zero-shot and generalized zero- shot remote sensing image scene classification

    Yansheng Li, Deyu Kong, Yongjun Zhang, Yihua Tan, and Ling Chen. Robust deep alignment network with remote sensing knowledge graph for zero-shot and generalized zero- shot remote sensing image scene classification. ISPRS Jour- nal of Photogrammetry and Remote Sensing , 179:145–158,

  13. [21]

    Learning deep representations for ground-to-aerial geolocal- ization

    Tsung-Yi Lin, Yin Cui, Serge Belongie, and James Hays. Learning deep representations for ground-to-aerial geolocal- ization. In Proceedings of the IEEE conference on computer vision and pattern recognition, pages 5007–5015, 2015. 3

    2015

  14. [22]

    Remoteclip: A vision lan- guage foundation model for remote sensing

    Fan Liu, Delong Chen, Zhangqingyun Guan, Xiaocong Zhou, Jiale Zhu, and Jun Zhou. Remoteclip: A vision lan- guage foundation model for remote sensing. arXiv preprint arXiv:2306.11029, 2023. 2, 12

    2023 arXiv

  15. [23]

    Decoupled weight decay regularization

    Ilya Loshchilov and Frank Hutter. Decoupled weight decay regularization. arXiv preprint arXiv:1711.05101, 2017. 5

    2017 arXiv

  16. [24]

    Jehanzeb Mirza, Leonid Karlinsky, Wei Lin, Ma- teusz Kozinski, Horst Possegger, Rogerio Feris, and Horst Bischof

    M. Jehanzeb Mirza, Leonid Karlinsky, Wei Lin, Ma- teusz Kozinski, Horst Possegger, Rogerio Feris, and Horst Bischof. Lafter: Label-free tuning of zero-shot classifier us- ing language and unlabeled image collections. InConference 9 on Neural Information Processing Systems (Neu...

    2023

  17. [25]

    Clip- cap: Clip prefix for image captioning

    Ron Mokady, Amir Hertz, and Amit H Bermano. Clip- cap: Clip prefix for image captioning. arXiv preprint arXiv:2111.09734, 2021. 6, 7

    2021 arXiv

  18. [26]

    Repre- sentation learning with contrastive predictive coding

    Aaron van den Oord, Yazhe Li, and Oriol Vinyals. Repre- sentation learning with contrastive predictive coding. arXiv preprint arXiv:1807.03748, 2018. 4

    2018 arXiv

  19. [27]

    Training language models to follow instructions with human feedback

    Long Ouyang, Jeffrey Wu, Xu Jiang, Diogo Almeida, Car- roll Wainwright, Pamela Mishkin, Chong Zhang, Sandhini Agarwal, Katarina Slama, Alex Ray, et al. Training language models to follow instructions with human feedback. Ad- vances in neural information processing systems, 35:...

    2022

  20. [28]

    What does a platypus look like? generating customized prompts for zero-shot image classification

    Sarah Pratt, Ian Covert, Rosanne Liu, and Ali Farhadi. What does a platypus look like? generating customized prompts for zero-shot image classification. In Proceedings of the IEEE/CVF International Conference on Computer Vision , pages 15691–15701, 2023. 2

    2023

  21. [29]

    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. In International conference on machine learning, ...

    2021

  22. [30]

    Remote sensing technology for mapping and monitoring land-cover and land-use change

    John Rogan and DongMei Chen. Remote sensing technology for mapping and monitoring land-cover and land-use change. Progress in planning, 61(4):301–325, 2004. 2

    2004

  23. [31]

    Waffling around for performance: Visual classification with random words and broad concepts

    Karsten Roth, Jae Myung Kim, A Koepke, Oriol Vinyals, Cordelia Schmid, and Zeynep Akata. Waffling around for performance: Visual classification with random words and broad concepts. arXiv preprint arXiv:2306.07282, 2023. 2, 3

    2023 arXiv

  24. [32]

    Meta-learning for few-shot land cover classification

    Marc Rußwurm, Sherrie Wang, Marco Korner, and David Lobell. Meta-learning for few-shot land cover classification. In Proceedings of the ieee/cvf conference on computer vision and pattern recognition workshops, pages 200–201, 2020. 2

    2020

  25. [33]

    Humans are poor few-shot classifiers for sentinel-2 land cover

    Marc Rußwurm, Sherrie Wang, and Devis Tuia. Humans are poor few-shot classifiers for sentinel-2 land cover. In IGARSS 2022-2022 IEEE International Geoscience and Re- mote Sensing Symposium, pages 4859–4862. IEEE, 2022. 2

    2022

  26. [34]

    Moein Shariatnia

    M. Moein Shariatnia. Simple CLIP, 4 2021. 3

    2021

  27. [35]

    Beyond cross-view image retrieval: Highly accurate vehicle localization using satel- lite image

    Yujiao Shi and Hongdong Li. Beyond cross-view image retrieval: Highly accurate vehicle localization using satel- lite image. In Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition, pages 17010– 17020, 2022. 3

    2022

  28. [36]

    Where am i looking at? joint location and orientation es- timation by cross-view matching

    Yujiao Shi, Xin Yu, Dylan Campbell, and Hongdong Li. Where am i looking at? joint location and orientation es- timation by cross-view matching. In Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition, pages 4064–4072, 2020. 3

    2020

  29. [37]

    mi- crosoft/planetarycomputer: October 2022, oct 2022

    Microsoft Open Source, Matt McFarland, Rob Emanuele, Dan Morris, and Tom Augspurger. mi- crosoft/planetarycomputer: October 2022, oct 2022. 3

    2022

  30. [38]

    Fine-grained landuse characterization us- ing ground-based pictures: a deep learning solution based on globally available data

    Shivangi Srivastava, John E Vargas Munoz, Sylvain Lobry, and Devis Tuia. Fine-grained landuse characterization us- ing ground-based pictures: a deep learning solution based on globally available data. International Journal of Geographi- cal Information Science, 34(6):1117–1136...

    2020

  31. [39]

    Fine-grained object recognition and zero-shot learn- ing in remote sensing imagery

    Gencer Sumbul, Ramazan Gokberk Cinbis, and Selim Ak- soy. Fine-grained object recognition and zero-shot learn- ing in remote sensing imagery. IEEE Transactions on Geo- science and Remote Sensing, 56(2):770–779, 2017. 2

    2017

  32. [40]

    Bigearthnet dataset with a new class-nomenclature for remote sensing image understanding

    Gencer Sumbul, Jian Kang, Tristan Kreuziger, Filipe Marcelino, Hugo Costa, Pedro Benevides, Mario Cae- tano, and Beg ¨um Demir. Bigearthnet dataset with a new class-nomenclature for remote sensing image understanding. arXiv preprint arXiv:2001.06372, 2020. 5

    2001 arXiv

  33. [41]

    The emergence of land change science for global environmen- tal change and sustainability

    Billie L Turner, Eric F Lambin, and Anette Reenberg. The emergence of land change science for global environmen- tal change and sustainability. Proceedings of the National Academy of Sciences, 104(52):20666–20671, 2007. 1

    2007

  34. [42]

    Skyscript: A large and seman- tically diverse vision-language dataset for remote sensing

    Zhecheng Wang, Rajanie Prabha, Tianyuan Huang, Jiajun Wu, and Ram Rajagopal. Skyscript: A large and seman- tically diverse vision-language dataset for remote sensing. In Proceedings of the AAAI Conference on Artificial Intel- ligence, volume 38, pages 5805–5813, 2024. 2, 6, 12

    2024

  35. [43]

    Mixed land use measurement and mapping with street view images and spatial context-aware prompts via zero-shot mul- timodal learning

    Meiliu Wu, Qunying Huang, Song Gao, and Zhou Zhang. Mixed land use measurement and mapping with street view images and spatial context-aware prompts via zero-shot mul- timodal learning. International Journal of Applied Earth Ob- servation and Geoinformation, 125:103591, 2023. 2

    2023

  36. [44]

    Concept-guided prompt learning for generalization in vision- language models

    Yi Zhang, Ce Zhang, Ke Yu, Yushun Tang, and Zhihai He. Concept-guided prompt learning for generalization in vision- language models. arXiv preprint arXiv:2401.07457, 2024. 2, 3

    2024 arXiv

  37. [45]

    Rs5m: A large scale vision-language dataset for remote sensing vision-language foundation model

    Zilun Zhang, Tiancheng Zhao, Yulong Guo, and Jianwei Yin. Rs5m: A large scale vision-language dataset for remote sensing vision-language foundation model. arXiv preprint arXiv:2306.11300, 2023. 2, 6, 12

    2023 arXiv

  38. [46]

    Conditional prompt learning for vision-language mod- els

    Kaiyang Zhou, Jingkang Yang, Chen Change Loy, and Ziwei Liu. Conditional prompt learning for vision-language mod- els. In Proceedings of the IEEE/CVF Conference on Com- puter Vision and Pattern Recognition , pages 16816–16825,

  39. [47]

    Learning to prompt for vision-language models

    Kaiyang Zhou, Jingkang Yang, Chen Change Loy, and Ziwei Liu. Learning to prompt for vision-language models. In- ternational Journal of Computer Vision, 130(9):2337–2348,

  40. [48]

    Transgeo: Trans- former is all you need for cross-view image geo-localization

    Sijie Zhu, Mubarak Shah, and Chen Chen. Transgeo: Trans- former is all you need for cross-view image geo-localization. In Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition, pages 1162–1171, 2022. 3

    2022

  41. [49]

    Vigor: Cross- view image geo-localization beyond one-to-one retrieval

    Sijie Zhu, Taojiannan Yang, and Chen Chen. Vigor: Cross- view image geo-localization beyond one-to-one retrieval. In Proceedings of the IEEE/CVF Conference on Computer Vi- sion and Pattern Recognition, pages 3640–3649, 2021. 3

    2021

  42. [50]

    Meta Prompt for Ground View Prompts

    Xiao Xiang Zhu, Devis Tuia, Lichao Mou, Gui-Song Xia, Liangpei Zhang, Feng Xu, and Friedrich Fraundorfer. Deep learning in remote sensing: A comprehensive review and list of resources. IEEE geoscience and remote sensing magazine, 5(4):8–36, 2017. 1, 2 10 A. Appendices A.1. Dat...

    2017

Pith tools

Reviewed August 11, 2026 · model on record in the stance chip above.