Pith. sign in

REVIEW 5 major objections 5 minor 63 references

Controllable Satellite-to-Street-View Synthesis with Precise Pose Alignment and Zero-Shot Environmental Control

T0 review · 5 major / 5 minor · reviewed 2026-08-09 · deepseek-v4-flash

Pith's one-line read A satellite-to-street-view diffusion pipeline uses an Iterative Homography Adjustment during denoising to keep generated ground images aligned with the satellite pose, and CLIP text guidance to control weather and lighting at inference.

desk verdict Integrates useful geometric conditioning for satellite-to-street-view diffusion, but the pose-alignment gains are marginal and the environmental control evaluation is partly circular; worth a careful review, not desk rejection. read the letter →

arxiv 2502.03498 v3 pith:M3HDMLV3 submitted 2025-02-05 eess.IV cs.GR

classification eess.IVcs.GR
keywords satellite-to-street-viewsynthesislatentdiffusionmodelposealignmentiterativehomographyadjustmentgeometriccross-attentionzero-shotenvironmentalcontroltext-guidedgenerationcross-viewlocalization
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

The paper tries to establish that a latent diffusion model can turn a satellite image into a geometrically consistent street-view panorama under a specified relative pose, while also letting the user choose weather and lighting at inference time. Its central claim is that pose errors, which ordinary diffusion ignores or encodes only implicitly, can be corrected by an Iterative Homography Adjustment that warps the latent code during DDIM denoising based on an InfoNCE alignment loss against satellite features. The paper also claims that environmental variation can be imposed zero-shot through CLIP text guidance on patches of the denoised image, without retraining or paired weather text in the dataset. If right, this would make satellite-to-street-view synthesis more usable for urban modeling and autonomous-driving data augmentation, because generated images would line up with the map and with the requested season or lighting.

What carries the argument

The load-bearing mechanism is the Iterative Homography Adjustment (IHA), a pose-correction loop inserted into DDIM sampling. At each early denoising step the model denoises the current latent to an estimate $z_{t,0}$, projects it onto an overhead view under a ground-plane assumption, computes an InfoNCE loss between ground-level features and satellite features at several candidate poses, and updates a homography matrix $H$ by gradient descent on that loss; the updated $H$ warps the latent via $z_t \otimes \mathrm{grid}(H)$. A second named component, Geometric Cross-Attention (GCA), replaces the usual cross-attention in the denoising U-Net: for each ground pixel it samples satellite features along $N=8$ reference height planes $h_i$ with learned offsets $\Delta h_i$ and softmax attention weights $A_i$, aggregating features as $\mathrm{GCA}(Q,V)=\sum_i A_i\,(V \otimes P(u_g,v_g,h_i+\Delta h_i))$. Together they inject explicit camera geometry into generation, while CLIP-text guidance supplies environmental control without retraining.

What would settle it

Take a test subset with tall buildings and strong parallax, generate the same scenes with and without IHA under identical noise, and measure lateral, longitudinal, and azimuth errors using a cross-view pose estimator; if IHA does not reduce those errors on that subset, the claim that a latent-space homography corrects pose is not supported.

Watch

Extended reading notes

Core claim

On the paper's own terms, the discovery is that both remaining bottlenecks in satellite-to-street-view synthesis—pose alignment and environmental diversity—can be handled at inference time rather than by new training data or heavy 3D reconstruction. The Geometric Cross-Attention module projects ground-pixel queries onto satellite feature maps through a set of reference height planes with learned offsets and attention weights, giving the diffusion model explicit geometric correspondences. The Iterative Homography Adjustment then computes a homography that best aligns the current denoised latent to the satellite at the requested camera pose, using a contrastive InfoNCE loss over candidate poses, and applies that homography to the latent at each of the first forty sampling steps. Environmental control is achieved separately by CLIP guidance: the denoised latent is split into patches, each patch is scored against a text prompt, and the gradient of that score steers sampling. The paper reports that this pipeline beats prior satellite-to-street-view generators on structural, perceptual, semantic, and depth similarity, and that the generated images improve cross-view localization when used as training augmentation.

Load-bearing premise

The method assumes that a single flat-plane image warp applied to the internal noise representation of the generated image can correct camera-pose error, which is exact only for flat ground and approximate for buildings and elevated structures.

Editorial extensions

If this is right

  • Street-view images generated from satellite imagery can keep a specified camera pose throughout generation, and pose errors appearing in the diffusion output are corrected during sampling rather than requiring retraining.
  • Environmental conditions such as seasons, weather, and lighting can be changed at inference time with text prompts, without paired weather-labelled training data.
  • The same framework works for panoramic ground images from CVUSA and VIGOR and for narrow-field-of-view camera images from KITTI.
  • Generated street-view data can serve as augmentation for training a cross-view vehicle localizer, and the paper reports it improves localization accuracy on KITTI.
  • Pixel-level metrics like RMSE and PSNR are not appropriate for comparing satellite-to-street-view synthesis; semantic, depth, and structural metrics should be used instead.

Reading between the lines

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

  • Because IHA operates purely on the latent code with a learned gradient on a homography, the same correction loop could in principle be attached to any latent diffusion model conditioned on a top-down reference; the paper only demonstrates this for its own finetuned U-Net, so this generalization is an inference, not a claim.
  • The paper's own limitations section admits that building facades and lane markings are imperfect, which points to the discrete-height-plus-homography geometry as the part most likely to bend under strong parallax; testing on high-rise scenes would show how far the claim extends.
  • The zero-shot environment control could likely be pushed toward finer attributes such as snow on roofs, wet asphalt, or sunset color temperature by using a stronger vision-language model or per-pixel CLIP features; the paper demonstrates only coarse season and weather prompts, so this is a forward extension rather than a demonstrated capability.
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

5 major / 5 minor

Summary. The paper proposes a latent diffusion framework for satellite-to-street-view synthesis. The two main contributions are an Iterative Homography Adjustment (IHA) scheme that updates a 3x3 homography in latent space during DDIM sampling to correct pose misalignment, and a Text-guided Zero-shot Environmental Control (ZoEC) strategy that uses CLIP-based classifier guidance to vary illumination and weather without retraining. A Geometric Cross-Attention (GCA) mechanism is also introduced to condition the diffusion model on satellite geometry. The method is evaluated on CVUSA, VIGOR, and KITTI with several image-quality metrics, a cross-view pose-estimation metric, and a data-augmentation experiment for cross-view localization.

Significance. If the method holds, it addresses a practically important problem: generating geometrically consistent street-view images from satellite inputs while allowing zero-shot control of environmental conditions. The paper has notable strengths: it evaluates pose alignment with an external cross-view pose estimator (CCVPE), it demonstrates a downstream application in data augmentation for cross-view localization (Appendix A.2), and it introduces a geometric conditioning mechanism that goes beyond implicit conditions. The central ideas are plausible and the experimental scope is substantial. However, the quantitative support for the pose-alignment claim is thin (small, partly inconsistent gains in Table 3), the environmental-control evaluation is partially circular because the reported CLIP score is the optimized objective, and the latent-space homography correction rests on an unverified equivariance assumption. These issues are fixable, but they affect load-bearing claims.

major comments (5)
  1. [§5.3, Table 3] The pose-alignment claim is supported by small and partly inconsistent differences: average distance changes from 10.74 to 10.51 with full pipeline (a 2.1% relative change), average angle from 17.60 to 17.14, but median angle worsens for LDM+IHA (6.64 to 6.73) and longitudinal @d=1 worsens for LDM+IHA (15.35 to 15.77). No error bars, significance tests, or multiple-seed variance are reported. Because the abstract claims that IHA 'effectively addresses pose misalignment,' Table 3 needs a stronger effect size or a proper statistical analysis to establish that the gains are not noise.
  2. [§4.2, Algorithm 1] IHA updates H by backpropagating the InfoNCE loss through the warped latent z_t ⊗ grid(H). For this correction to transfer to the decoded street-view image, the latent space must be approximately equivariant to image-space homographies, i.e., D(H·z) ≈ H·D(z). No evidence for this equivariance is provided. If it fails, the gradient of L_pose with respect to H may reduce the latent loss without correctly fixing the pose of the decoded image. Please add a direct equivariance check (comparing D(warp(z)) and warp(D(z)) on a set of generated samples) and/or evaluate pose metrics on the decoded images after applying IHA in latent space versus after applying H in image space.
  3. [§4.3, Table 4] The environmental-control evaluation is partially circular. L_text in Eq. (8) is defined as 1 minus the mean CLIP cosine similarity between patches of z_{t,0} and the text c_text, and Table 4 reports the same CLIP cosine similarity as 'Clip Score' as evidence of environmental alignment. This score is an optimization objective, not an independent metric. The BLIP score is a useful step, but the authors should report at least one additional metric that is not directly optimized, such as a human evaluation, a separate vision-language model, or a classifier trained on weather/season labels.
  4. [Appendix A.4] The derivation of the guided-sampling formula contains a variable mismatch: Eq. (11) writes p(z_{t-1}|g_pose, g_text) = C p(g_pose, g_text | z_t) p(z_t), mixing z_{t-1} on the left with z_t on the right. The subsequent independence assumption p(g_pose, g_text | z_t) = p(g_pose | z_t) p(g_text | z_t) is also unstated and does not generally hold if pose and environment are correlated in the image. Please correct the derivation to a consistent Bayes update and explicitly state the independence approximation.
  5. [§4.2, last paragraph] The sentence 'During the inference stage, both the denoising Unet and the Homography Adjustment network are trained' contradicts the described inference-time optimization of H and the earlier statement that the correction is applied through H updates. If the Homography Adjustment network's weights are actually trained at inference, the paper must explain where the supervision comes from and how this is consistent with the zero-shot description; otherwise the method cannot be reproduced from the text.
minor comments (5)
  1. [§5.1] The height hypotheses list {−3,−2,−1,1,2,3,4,5} is given without a sign convention relative to the ground plane or a sensitivity study; the sentence 'This constitutes an empirical setup' is too vague to guide reimplementation.
  2. [Algorithm 1] The comment 'H is a diagonal matrix of ones' is ambiguous or incorrect: a 3x3 homography initialized for identity should be the identity matrix, not a diagonal matrix of ones (which would have ones on the diagonal and zeros elsewhere; that is the identity only if the off-diagonal zeros are intended). Please clarify.
  3. [Figure 2] The notation in Figure 2 is cluttered: the labels 'each patch', 'Pose Correction', and 'Env Control' are difficult to read, and the relationship between the 'Warp' boxes and the algorithm steps is unclear.
  4. [References] Rombach et al. 2022a and 2022b are the same CVPR paper; please consolidate the duplicate reference.
  5. [Reproducibility] The paper does not state whether code or trained models will be released; sharing code would significantly help reproducibility given the number of hyperparameters (γ, λ, τ, the height set, the IHA window, and the patch partition).

Circularity Check

1 steps flagged · score 4.0 of 10

Partial circularity in ZoEC evaluation: Eq. 8's CLIP guidance objective is reported as Table 4's Clip Score; pose alignment is independently evaluated.

  1. fitted input called prediction [Sec. 4.3 (Eq. 8); Sec. 4.4 (text-similarity metric); Table 4]
    ""Ltext(zt, ctext) = 1− 1/N Σ sim(CLIP(z^l_{t,0}), CLIP(ctext))" (Eq. 8). "Finally, for text similarity, we compute the similarity of Clip Radford et al. (2021) features for evaluation" (Sec. 4.4). "Table 4: The similarity between generated images and environmental text prompt.""

    The ZoEC guidance in Eq. 8 directly maximizes CLIP cosine similarity between generated latent patches and the text prompt. The headline quantitative evidence for environmental control, Table 4's Clip Score, is the same CLIP cosine similarity evaluated on the generated images. Therefore the high Clip Score is partly forced by the sampling objective rather than by independent verification that the scene actually displays the requested season. The Blip Score and qualitative figures are separate evidence, so the circularity is partial and confined to the CLIP-based evaluation.

full rationale

Score 4 reflects one partial circularity. The paper's pose-alignment claim is not circular: IHA adjusts a latent homography by minimizing an InfoNCE pose loss, but the Table 3 evaluation uses the external CCVPE pose estimator, so the reported pose metrics are not the training objective. The GCA projection and the Appendix A.4 classifier-guidance derivation are standard reformulations with stated assumptions and do not smuggle in the target result. The self-citations to Shi et al. are for projection/localization components and are not used to forbid alternatives or justify a uniqueness claim. The remaining issue is the environmental-control evaluation: Eq. 8 optimizes CLIP cosine similarity and Table 4 reports CLIP cosine similarity, so that part of the evidence reduces by construction. This does not invalidate the whole paper because qualitative results and the BLIP recall metric are independent, and the pose half is externally validated; hence 4 rather than 6+.

Assumptions & free parameters 6 free parameters · 4 assumptions · 0 invented entities

The ledger reflects that the system builds on stable diffusion, conventional cross-view datasets with known poses, and frozen CLIP. No physical entity is introduced; the new elements are trainable modules. The main unpaid premises are the piecewise-planar height representation, the latent-space homography assumption, and the use of CLIP as the environmental reward.

free parameters (6)
  • GCA reference height hypotheses h_i = {-3, -2, -1, 1, 2, 3, 4, 5}, N=8
    Set empirically in Sec. 5.1; determines which satellite features are attended to by Eq. 5.
  • Text guidance strength gamma = not reported
    Eq. 4 and Eq. 14 weight the CLIP gradient; the value is not given.
  • Pose guidance strength lambda = not reported
    Eq. 4 and Eq. 14 weight the pose gradient; the value is not given.
  • InfoNCE temperature tau = not reported
    Eq. 6 uses tau in the contrastive loss; no value or sensitivity analysis is reported.
  • IHA application window = first 40 of 50 DDIM steps
    Sec. 5.1 states this fixed window without an ablation.
  • Text loss patch partition = not reported
    Eq. 8 partitions z_{t-1} into N patches; the patch size and number are unspecified.
assumptions (4)
  • standard math Bayes factorization and independence of pose and text guidance in Appendix A.4
    Eq. 11 assumes p(g_pose, g_text | z_t) = p(g_pose | z_t) p(g_text | z_t); this independence is asserted, not argued.
  • domain assumption Discrete height-plane scene representation with learned offsets captures 3D geometry
    Eq. 5 projects each ground pixel through hypothesized heights; no estimated depth is used, so the geometric link between views depends on this model.
  • ad hoc to paper Homography warping of the latent z_t approximates image-space homography correction
    Algorithm 1 and Sec. 4.2 apply H to latent grids; the equivalence to image-space warp is assumed without justification.
  • domain assumption CLIP cosine similarity is a sufficient reward for weather and season control
    Eq. 8 uses frozen CLIP as the only environmental supervision; if CLIP does not separate seasons or weather well, the control fails.

how reviews work

0 comments
Cite this review

Pith. "Pith review of Controllable Satellite-to-Street-View Synthesis with Precise Pose Alignment and Zero-Shot Environmental Control." pith.science (2026). https://pith.science/paper/M3HDMLV3

@misc{pith2026250203498,
  author       = {Pith},
  title        = {Pith review of: Controllable Satellite-to-Street-View Synthesis with Precise Pose Alignment and Zero-Shot Environmental Control},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/M3HDMLV3}},
  note         = {Machine review of arXiv:2502.03498}
}
read the original abstract

Generating street-view images from satellite imagery is a challenging task, particularly in maintaining accurate pose alignment and incorporating diverse environmental conditions. While diffusion models have shown promise in generative tasks, their ability to maintain strict pose alignment throughout the diffusion process is limited. In this paper, we propose a novel Iterative Homography Adjustment (IHA) scheme applied during the denoising process, which effectively addresses pose misalignment and ensures spatial consistency in the generated street-view images. Additionally, currently, available datasets for satellite-to-street-view generation are limited in their diversity of illumination and weather conditions, thereby restricting the generalizability of the generated outputs. To mitigate this, we introduce a text-guided illumination and weather-controlled sampling strategy that enables fine-grained control over the environmental factors. Extensive quantitative and qualitative evaluations demonstrate that our approach significantly improves pose accuracy and enhances the diversity and realism of generated street-view images, setting a new benchmark for satellite-to-street-view generation tasks.

Figures

Figures reproduced from arXiv: 2502.03498 by the authors.

Figure 1
Figure 1. Our method synthesizes ground-level images from satellite maps by integrating Iterative Homogra￾phy Adjustment (IHA) to refine geometric alignment and Text-guided Zero-shot Environmental Control (ZoEC) to provide flexible environmental control, enabling precise pose alignment and diverse scene generation. As shown in [PITH_FULL_IMAGE:figures/full_fig_p002_1.png] view at source ↗
Figure 2
Figure 2. Overview of our framework. Our approach aims to utilize satellite images to generate corresponding ground images. We leverage geometric relationships extensively and have the capability to alter the features of the generated images based on different text prompts. where σt controls the sampling’s stochasticity, and µθ(zt, t, c) is calculated as: µθ(zt, t, c) = √ α¯t−1  zt − √ 1 − α¯t · ϵθ(zt, t, c) √ α¯t  + q 1 − … view at source ↗
Figure 3
Figure 3. Homography Adjustment. We compare ground-level images with satellite images, calculating the loss by assessing the relative pose of the ground images at time t against a predefined pose. Ground images are then adjusted based on this comparison. 4.1 CROSS-VIEW CONDITIONING MECHANISM Instead of implicitly encoding global feature vectors of posed satellite images, we propose a Geo￾metric Cross-Attention (GCA) mechanism… view at source ↗
Figures from the paper (14 more)
Figure 4
Figure 4. Figure 4: Qualitative comparison with previous work on the CVUSA dataset, our framework is able to main￾tain better geometric relationships [PITH_FULL_IMAGE:figures/full_fig_p007_4.png]
Figure 5
Figure 5. Figure 5: Qualitative comparison with previous work on the VIGOR dataset, our model can effectively capture road surface information from satellite images, resulting in clearer and more distinct road lines [PITH_FULL_IMAGE:figures/full_fig_p008_5.png]
Figure 6
Figure 6. Figure 6: Qualitative ablation experiments of IHA on KITTI and CVUSA datasets. IHA can correct pose offsets in the inference process effectively [PITH_FULL_IMAGE:figures/full_fig_p009_6.png]
Figure 7
Figure 7. Figure 7: Qualitative comparison with ControlNet on CVUSA. ControlNet trained under weak textual condi￾tions fails to effectively adjust image information based on text prompts. Our Zero-shot Environmental Control can adjust image information based on text prompts while preservi…
Figure 8
Figure 8. Figure 8: In the first row, we demonstrate that ground images generated from satellite maps exhibit varying offsets under different initial noise conditions. The second row illustrates the pose correction results of the first row using the Iterative Homography Adjustment (IHA). …
Figure 9
Figure 9. Figure 9: The relationship between Homography Adjustment and the camera position in satellite imagery. The corresponding satellite position and ground images are labeled with arrows and borders using the same colors. The Homography transformation is defined by a 3x3 matrix and c…
Figure 10
Figure 10. Figure 10: The Homography Adjustment operation in the DDIM process. The visual results of each step stem from D(zt,0) . Due to variations in initial noise and uncertainties in the implicit learning of features, diffusion does not always yield favorable outcomes. As depicted in …
Figure 11
Figure 11. Figure 11: Qualitative visualization for generated images using different methods in complex scenes. encode the buildings due to few occupant pixels. Owing to the geometric reasoning capabilities of GCA and IHA, our method excels at mining deep semantic features at the correct c…
Figure 12
Figure 12. Figure 12: Visualization results of the GCA mechanism. The feature maps are visualized using PCA, while the attention maps are displayed with a jet color mapping. In the middle row, we present ground feature maps sampled from satellite features according to four candidate height…
Figure 13
Figure 13. Figure 13: Using LLAVA-generated text for ground images can correspond to multiple ground pictures, while satellite images correspond one-to-one with ground images. the weight assigned to the satellite image condition far exceeded that of the environmental descrip￾tion condition…
Figure 14
Figure 14. Figure 14: Controllable generation in the Chinese context. This figure displays generated images of spring, autumn, and winter from top to bottom. In order to ensure compatibility across linguistic contexts, our framework provides a flexible inter￾face that can integrate with va…
Figure 15
Figure 15. Figure 15: Generating under different textual conditions to obtain text-guided image results while maintaining the structure of the images A.11 MORE RESULTS FROM POSE ALIGNMENT. (a) Sat (b) LDM (c) GT (d) LDM W. IHA [PITH_FULL_IMAGE:figures/full_fig_p021_15.png]
Figure 16
Figure 16. Figure 16: Results of IHA on the KITTI dataset. 21 [PITH_FULL_IMAGE:figures/full_fig_p021_16.png]
Figure 17
Figure 17. Figure 17: Results of IHA on the CVUSA dataset. 22 [PITH_FULL_IMAGE:figures/full_fig_p022_17.png]

Discussion (0). Continue with ORCID to comment.

Reference graph

Works this paper leans on

63 extracted references · 40 canonical work pages

  1. [1]

    A-bdd: Leveraging data augmentations for safe autonomous driving in adverse weather and lighting

    Felix Assion, Florens Gressner, Nitin Augustine, Jona Klemenc, Ahmed Hammam, Alexandre Krattinger, Holger Trittenbach, and Sascha Riemer. A-bdd: Leveraging data augmentations for safe autonomous driving in adverse weather and lighting. arXiv preprint arXiv:2408.06071, 2024

  2. [2]

    Blended diffusion for text-driven editing of natural images

    Omri Avrahami, Dani Lischinski, and Ohad Fried. Blended diffusion for text-driven editing of natural images. In Proceedings of the IEEE/CVF conference on computer vision and pattern recognition, pp.\ 18208--18218, 2022

  3. [3]

    Instructpix2pix: Learning to follow image editing instructions

    Tim Brooks, Aleksander Holynski, and Alexei A Efros. Instructpix2pix: Learning to follow image editing instructions. In Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition, pp.\ 18392--18402, 2023

  4. [4]

    Emerging properties in self-supervised vision transformers

    Mathilde Caron, Hugo Touvron, Ishan Misra, Herv \'e J \'e gou, Julien Mairal, Piotr Bojanowski, and Armand Joulin. Emerging properties in self-supervised vision transformers. In Proceedings of the IEEE/CVF international conference on computer vision, pp.\ 9650--9660, 2021

  5. [5]

    Tac-gan-text conditioned auxiliary classifier generative adversarial network

    Ayushman Dash, John Cristian Borges Gamboa, Sheraz Ahmed, Marcus Liwicki, and Muhammad Zeshan Afzal. Tac-gan-text conditioned auxiliary classifier generative adversarial network. arXiv preprint arXiv:1703.06412, 2017

  6. [6]

    Diffusion models beat gans on image synthesis

    Prafulla Dhariwal and Alexander Nichol. Diffusion models beat gans on image synthesis. Advances in neural information processing systems, 34: 0 8780--8794, 2021

  7. [7]

    Scp-diff: Photo-realistic semantic image synthesis with spatial-categorical joint prior

    Huan-ang Gao, Mingju Gao, Jiaju Li, Wenyi Li, Rong Zhi, Hao Tang, and Hao Zhao. Scp-diff: Photo-realistic semantic image synthesis with spatial-categorical joint prior. arXiv preprint arXiv:2403.09638, 2024 a

  8. [8]

    Cat3d: Create anything in 3d with multi-view diffusion models

    Ruiqi Gao, Aleksander Holynski, Philipp Henzler, Arthur Brussee, Ricardo Martin-Brualla, Pratul Srinivasan, Jonathan T Barron, and Ben Poole. Cat3d: Create anything in 3d with multi-view diffusion models. arXiv preprint arXiv:2405.10314, 2024 b

Show all 63 references
  1. [9]

    MagicDrive : Street view generation with diverse 3d geometry control

    Ruiyuan Gao, Kai Chen, Enze Xie, Lanqing Hong, Zhenguo Li, Dit-Yan Yeung, and Qiang Xu. MagicDrive : Street view generation with diverse 3d geometry control. In International Conference on Learning Representations, 2024 c

  2. [10]

    Vision meets robotics: The kitti dataset

    Andreas Geiger, Philip Lenz, Christoph Stiller, and Raquel Urtasun. Vision meets robotics: The kitti dataset. The International Journal of Robotics Research, 32 0 (11): 0 1231--1237, 2013

  3. [11]

    Squeezenet: Alexnet-level accuracy with 50x fewer parameters and< 0.5 mb model size

    Forrest N Iandola. Squeezenet: Alexnet-level accuracy with 50x fewer parameters and< 0.5 mb model size. arXiv preprint arXiv:1602.07360, 2016

  4. [12]

    Image-to-image translation with conditional adversarial networks

    Phillip Isola, Jun-Yan Zhu, Tinghui Zhou, and Alexei A Efros. Image-to-image translation with conditional adversarial networks. CVPR, 2017

  5. [13]

    Segment anything

    Alexander Kirillov, Eric Mintun, Nikhila Ravi, Hanzi Mao, Chloe Rolland, Laura Gustafson, Tete Xiao, Spencer Whitehead, Alexander C Berg, Wan-Yen Lo, et al. Segment anything. In Proceedings of the IEEE/CVF International Conference on Computer Vision, pp.\ 4015--4026, 2023

  6. [14]

    Imagenet classification with deep convolutional neural networks

    Alex Krizhevsky, Ilya Sutskever, and Geoffrey E Hinton. Imagenet classification with deep convolutional neural networks. Communications of the ACM, 60 0 (6): 0 84--90, 2017

  7. [15]

    Slicematch: Geometry-guided aggregation for cross-view pose estimation

    Ted de Vries Lentsch, Zimin Xia, Holger Caesar, and Julian FP Kooij. Slicematch: Geometry-guided aggregation for cross-view pose estimation. arXiv preprint arXiv:2211.14651, 2022

  8. [16]

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

    Junnan Li, Dongxu Li, Caiming Xiong, and Steven Hoi. Blip: Bootstrapping language-image pre-training for unified vision-language understanding and generation. In International conference on machine learning, pp.\ 12888--12900. PMLR, 2022

  9. [17]

    Crossviewdiff: A cross-view diffusion model for satellite-to-street view synthesis

    Weijia Li, Jun He, Junyan Ye, Huaping Zhong, Zhimeng Zheng, Zilong Huang, Dahua Lin, and Conghui He. Crossviewdiff: A cross-view diffusion model for satellite-to-street view synthesis. arXiv preprint arXiv:2408.14765, 2024 a

  10. [18]

    Drivingdiffusion: Layout-guided multi-view driving scene video generation with latent diffusion model

    Xiaofan Li, Yifu Zhang, and Xiaoqing Ye. Drivingdiffusion: Layout-guided multi-view driving scene video generation with latent diffusion model. arXiv preprint arXiv:2310.07771, 2023 a

  11. [19]

    Gligen: Open-set grounded text-to-image generation

    Yuheng Li, Haotian Liu, Qingyang Wu, Fangzhou Mu, Jianwei Yang, Jianfeng Gao, Chunyuan Li, and Yong Jae Lee. Gligen: Open-set grounded text-to-image generation. In Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition, pp.\ 22511--22521, 2023 b

  12. [20]

    Sat2vid: Street-view panoramic video synthesis from a single satellite image

    Zuoyue Li, Zhenqiang Li, Zhaopeng Cui, Rongjun Qin, Marc Pollefeys, and Martin R Oswald. Sat2vid: Street-view panoramic video synthesis from a single satellite image. In Proceedings of the IEEE/CVF International Conference on Computer Vision, pp.\ 12436--12445, 2021

  13. [21]

    Sat2scene: 3d urban scene generation from satellite images with diffusion

    Zuoyue Li, Zhenqiang Li, Zhaopeng Cui, Marc Pollefeys, and Martin R Oswald. Sat2scene: 3d urban scene generation from satellite images with diffusion. In Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition, pp.\ 7141--7150, 2024 b

  14. [22]

    Visual instruction tuning

    Haotian Liu, Chunyuan Li, Qingyang Wu, and Yong Jae Lee. Visual instruction tuning. Advances in neural information processing systems, 36, 2024

  15. [23]

    Zero-1-to-3: Zero-shot one image to 3d object

    Ruoshi Liu, Rundi Wu, Basile Van Hoorick, Pavel Tokmakov, Sergey Zakharov, and Carl Vondrick. Zero-1-to-3: Zero-shot one image to 3d object. In Proceedings of the IEEE/CVF international conference on computer vision, pp.\ 9298--9309, 2023

  16. [24]

    Geometry-aware satellite-to-ground image synthesis for urban areas

    Xiaohu Lu, Zuoyue Li, Zhaopeng Cui, Martin R Oswald, Marc Pollefeys, and Rongjun Qin. Geometry-aware satellite-to-ground image synthesis for urban areas. In Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition, pp.\ 859--867, 2020

  17. [25]

    Im-3d: Iterative multiview diffusion and reconstruction for high-quality 3d generation

    Luke Melas-Kyriazi, Iro Laina, Christian Rupprecht, Natalia Neverova, Andrea Vedaldi, Oran Gafni, and Filippos Kokkinos. Im-3d: Iterative multiview diffusion and reconstruction for high-quality 3d generation. arXiv preprint arXiv:2402.08682, 2024

  18. [26]

    Representation learning with contrastive predictive coding

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

  19. [27]

    Dreamfusion: Text-to-3d using 2d diffusion

    Ben Poole, Ajay Jain, Jonathan T Barron, and Ben Mildenhall. Dreamfusion: Text-to-3d using 2d diffusion. arXiv preprint arXiv:2209.14988, 2022

  20. [28]

    Sat2density: Faithful density learning from satellite-ground image pairs

    Ming Qian, Jincheng Xiong, Gui-Song Xia, and Nan Xue. Sat2density: Faithful density learning from satellite-ground image pairs. In Proceedings of the IEEE/CVF International Conference on Computer Vision, pp.\ 3683--3692, 2023

  21. [29]

    Learning transferable visual models from natural language supervision

    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 supervision. In International conference on machine learning, pp...

  22. [30]

    Cross-view image synthesis using conditional gans

    Krishna Regmi and Ali Borji. Cross-view image synthesis using conditional gans. In Proceedings of the IEEE conference on Computer Vision and Pattern Recognition, pp.\ 3501--3510, 2018

  23. [31]

    High-resolution image synthesis with latent diffusion models

    Robin Rombach, Andreas Blattmann, Dominik Lorenz, Patrick Esser, and Bj\"orn Ommer. High-resolution image synthesis with latent diffusion models. In Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR), pp.\ 10684--10695, June 2022 a

  24. [32]

    High-resolution image synthesis with latent diffusion models

    Robin Rombach, Andreas Blattmann, Dominik Lorenz, Patrick Esser, and Bj \"o rn Ommer. High-resolution image synthesis with latent diffusion models. In Proceedings of the IEEE/CVF conference on computer vision and pattern recognition, pp.\ 10684--10695, 2022 b

  25. [33]

    Dae-gan: Dynamic aspect-aware gan for text-to-image synthesis

    Shulan Ruan, Yong Zhang, Kun Zhang, Yanbo Fan, Fan Tang, Qi Liu, and Enhong Chen. Dae-gan: Dynamic aspect-aware gan for text-to-image synthesis. In Proceedings of the IEEE/CVF international conference on computer vision, pp.\ 13960--13969, 2021

  26. [34]

    Geosynth: Contextually-aware high-resolution satellite image synthesis

    Srikumar Sastry, Subash Khanal, Aayush Dhakal, and Nathan Jacobs. Geosynth: Contextually-aware high-resolution satellite image synthesis. In Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition, pp.\ 460--470, 2024

  27. [35]

    Matching local self-similarities across images and videos

    Eli Shechtman and Michal Irani. Matching local self-similarities across images and videos. In 2007 IEEE conference on computer vision and pattern recognition, pp.\ 1--8. IEEE, 2007

  28. [36]

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

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

  29. [37]

    Spatial-aware feature aggregation for image based cross-view geo-localization

    Yujiao Shi, Liu Liu, Xin Yu, and Hongdong Li. Spatial-aware feature aggregation for image based cross-view geo-localization. Advances in Neural Information Processing Systems, 32, 2019

  30. [38]

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

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

  31. [39]

    Optimal feature transport for cross-view image geo-localization

    Yujiao Shi, Xin Yu, Liu Liu, Tong Zhang, and Hongdong Li. Optimal feature transport for cross-view image geo-localization. In Proceedings of the AAAI Conference on Artificial Intelligence, volume 34, pp.\ 11990--11997, 2020 b

  32. [40]

    Geometry-guided street-view panorama synthesis from satellite imagery

    Yujiao Shi, Dylan Campbell, Xin Yu, and Hongdong Li. Geometry-guided street-view panorama synthesis from satellite imagery. IEEE Transactions on Pattern Analysis and Machine Intelligence, 44 0 (12): 0 10009--10022, 2022 a

  33. [41]

    Accurate 3-dof camera geo-localization via ground-to-satellite image matching

    Yujiao Shi, Xin Yu, Liu Liu, Dylan Campbell, Piotr Koniusz, and Hongdong Li. Accurate 3-dof camera geo-localization via ground-to-satellite image matching. IEEE transactions on pattern analysis and machine intelligence, 45 0 (3): 0 2682--2697, 2022 b

  34. [42]

    Cvlnet: Cross-view semantic correspondence learning for video-based camera localization

    Yujiao Shi, Xin Yu, Shan Wang, and Hongdong Li. Cvlnet: Cross-view semantic correspondence learning for video-based camera localization. In Asian Conference on Computer Vision, pp.\ 123--141. Springer, 2022 c

  35. [43]

    Boosting 3-dof ground-to-satellite camera localization accuracy via geometry-guided cross-view transformer

    Yujiao Shi, Fei Wu, Akhil Perincherry, Ankit Vora, and Hongdong Li. Boosting 3-dof ground-to-satellite camera localization accuracy via geometry-guided cross-view transformer. In Proceedings of the IEEE/CVF International Conference on Computer Vision, pp.\ 21516--21526, 2023

  36. [44]

    Weakly-supervised camera localization by ground-to-satellite image registration

    Yujiao Shi, Hongdong Li, Akhil Perincherry, and Ankit Vora. Weakly-supervised camera localization by ground-to-satellite image registration. In European Conference on Computer Vision, pp.\ 39--57. Springer, 2025

  37. [45]

    Score-based generative modeling through stochastic differential equations

    Yang Song, Jascha Sohl-Dickstein, Diederik P Kingma, Abhishek Kumar, Stefano Ermon, and Ben Poole. Score-based generative modeling through stochastic differential equations. arXiv preprint arXiv:2011.13456, 2020

  38. [46]

    Learning dense flow field for highly-accurate cross-view camera localization

    Zhenbo Song, Jianfeng Lu, Yujiao Shi, et al. Learning dense flow field for highly-accurate cross-view camera localization. Advances in Neural Information Processing Systems, 36, 2024

  39. [47]

    Df-gan: A simple and effective baseline for text-to-image synthesis

    Ming Tao, Hao Tang, Fei Wu, Xiao-Yuan Jing, Bing-Kun Bao, and Changsheng Xu. Df-gan: A simple and effective baseline for text-to-image synthesis. In Proceedings of the IEEE/CVF conference on computer vision and pattern recognition, pp.\ 16515--16525, 2022

  40. [48]

    Splicing vit features for semantic appearance transfer

    Narek Tumanyan, Omer Bar-Tal, Shai Bagon, and Tali Dekel. Splicing vit features for semantic appearance transfer. In Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition, pp.\ 10748--10757, 2022

  41. [49]

    Cross-view panorama image synthesis

    Songsong Wu, Hao Tang, Xiao-Yuan Jing, Haifeng Zhao, Jianjun Qian, Nicu Sebe, and Yan Yan. Cross-view panorama image synthesis. IEEE Transactions on Multimedia, 25: 0 3546--3559, 2022

  42. [50]

    Visual cross-view metric localization with dense uncertainty estimates

    Zimin Xia, Olaf Booij, Marco Manfredi, and Julian FP Kooij. Visual cross-view metric localization with dense uncertainty estimates. In European Conference on Computer Vision, pp.\ 90--106. Springer, 2022

  43. [51]

    Convolutional cross-view pose estimation

    Zimin Xia, Olaf Booij, and Julian FP Kooij. Convolutional cross-view pose estimation. IEEE Transactions on Pattern Analysis and Machine Intelligence, 2023

  44. [52]

    Adapting fine-grained cross-view localization to areas without fine ground truth

    Zimin Xia, Yujiao Shi, Hongdong Li, and Julian FP Kooij. Adapting fine-grained cross-view localization to areas without fine ground truth. In European Conference on Computer Vision, pp.\ 397--415. Springer, 2025

  45. [53]

    Geospecific view generation--geometry-context aware high-resolution ground view inference from satellite views

    Ningli Xu and Rongjun Qin. Geospecific view generation--geometry-context aware high-resolution ground view inference from satellite views. arXiv preprint arXiv:2407.08061, 2024

  46. [54]

    Chinese clip: Contrastive vision-language pretraining in chinese

    An Yang, Junshu Pan, Junyang Lin, Rui Men, Yichang Zhang, Jingren Zhou, and Chang Zhou. Chinese clip: Contrastive vision-language pretraining in chinese. arXiv preprint arXiv:2211.01335, 2022

  47. [55]

    Depth anything: Unleashing the power of large-scale unlabeled data

    Lihe Yang, Bingyi Kang, Zilong Huang, Xiaogang Xu, Jiashi Feng, and Hengshuang Zhao. Depth anything: Unleashing the power of large-scale unlabeled data. In Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition, pp.\ 10371--10381, 2024

  48. [56]

    Predicting ground-level scene layout from aerial imagery

    Menghua Zhai, Zachary Bessinger, Scott Workman, and Nathan Jacobs. Predicting ground-level scene layout from aerial imagery. In Proceedings of the IEEE Conference on Computer Vision and Pattern Recognition, pp.\ 867--875, 2017

  49. [57]

    Adding conditional control to text-to-image diffusion models

    Lvmin Zhang, Anyi Rao, and Maneesh Agrawala. Adding conditional control to text-to-image diffusion models. In Proceedings of the IEEE/CVF International Conference on Computer Vision, pp.\ 3836--3847, 2023

  50. [58]

    The unreasonable effectiveness of deep features as a perceptual metric

    Richard Zhang, Phillip Isola, Alexei A Efros, Eli Shechtman, and Oliver Wang. The unreasonable effectiveness of deep features as a perceptual metric. In Proceedings of the IEEE conference on computer vision and pattern recognition, pp.\ 586--595, 2018

  51. [59]

    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 Vision and Pattern Recognition, pp.\ 3640--3649, 2021

  52. [60]

    write newline

    " write newline "" before.all 'output.state := FUNCTION n.dashify 't := "" t empty not t #1 #1 substring "-" = t #1 #2 substring "--" = not "--" * t #2 global.max substring 't := t #1 #1 substring "-" = "-" * t #2 global.max substring 't := while if t #1 #1 substring * t #2 gl...

  53. [61]

    @esa (Ref

    \@ifxundefined[1] #1\@undefined \@firstoftwo \@secondoftwo \@ifnum[1] #1 \@firstoftwo \@secondoftwo \@ifx[1] #1 \@firstoftwo \@secondoftwo [2] @ #1 \@temptokena #2 #1 @ \@temptokena \@ifclassloaded agu2001 natbib The agu2001 class already includes natbib coding, so you should ...

  54. [62]

    \@lbibitem[] @bibitem@first@sw\@secondoftwo \@lbibitem[#1]#2 \@extra@b@citeb \@ifundefined br@#2\@extra@b@citeb \@namedef br@#2 \@nameuse br@#2\@extra@b@citeb \@ifundefined b@#2\@extra@b@citeb @num @parse #2 @tmp #1 NAT@b@open@#2 NAT@b@shut@#2 \@ifnum @merge>\@ne @bibitem@firs...

  55. [63]

    @open @close @open @close and [1] URL: #1 \@ifundefined chapter * \@mkboth \@ifxundefined @sectionbib * \@mkboth * \@mkboth\@gobbletwo \@ifclassloaded amsart * \@ifclassloaded amsbook * \@ifxundefined @heading @heading NAT@ctr thebibliography [1] @ \@biblabel @NAT@ctr \@bibset...

Pith tools

Reviewed August 9, 2026 · model on record in the stance chip above.