Pith. sign in

REVIEW 4 major objections 4 minor 75 references

Adapting Vehicle Detectors for Aerial Imagery to Unseen Domains with Weak Supervision

T0 review · 4 major / 4 minor · reviewed 2026-08-06 · deepseek-v4-flash

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

Pith's one-line read A diffusion-based pipeline generates synthetic aerial images and derives their labels from stacked cross-attention maps, lifting cross-domain vehicle detection to 75.7 AP50 without target bounding-box annotations.

desk verdict Novel diffusion-based cross-domain labeling pipeline with strong reported gains, but the key transfer assumption is never directly measured and no artifacts are released. read the letter →

arxiv 2507.20976 v1 pith:XWVDJ62Z submitted 2025-07-28 cs.CV

Xiao Fang , Minhyek Jeon , Zheyang Qin , Stanislav Panev , Celso de Melo , Shuowen Hu , Shayok Chakraborty , Fernando De la Torre This is my paper · ORCID
classification cs.CV
keywords cross-domainobjectdetectionaerialvehiclelatentdiffusionmodelscross-attentionmapsweaklysuperviseddomainadaptationsyntheticdataaugmentationsmallshift
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 claims that you can adapt a vehicle detector to a new aerial domain without bounding-box annotations by generating synthetic images of the target domain and labeling them automatically. The labels come from cross-attention maps inside a fine-tuned text-to-image diffusion model: three stacked maps for a car concept, a learned foreground token, and a learned background token are turned into detector inputs, and a detector trained on synthetic source attention maps transfers to synthetic target attention maps. With only image-level labels for the target, the pipeline then trains the final detector on synthetic target images and their pseudo-labels. Across two domain shifts, this method reports AP50 gains over source-only training of 4–23%, and best scores of 75.4% and 75.7% with YOLOv8, beating weakly supervised and unsupervised adaptation baselines. The paper also contributes two new annotated aerial datasets, LINZ and UGRC.

What carries the argument

The load-bearing object is the stacked (multi-channel) cross-attention map of a fine-tuned latent diffusion model. For each synthetic image, the paper extracts cross-attention maps for the word 'car', for a learnable foreground token [V1], and for a learnable background token [V2]/[V3], normalizes them, and stacks them into a three-channel 'style-less' representation. These maps are what lets labels travel from source to target: a detector trained on stacked maps of labeled synthetic source images is applied to stacked maps of unlabeled synthetic target images, producing boxes that are then attached to the corresponding synthetic target RGB images. Two fine-tuning stages align the diffusion model with aerial views and with the learned tokens, using a total-variation loss that pushes the foreground token map toward the category map and the background token maps toward its complement.

What would settle it

Train a binary classifier to distinguish stacked source attention maps from stacked target attention maps; if the classifier performs well above chance, the attention maps carry domain information and the label-transfer chain is suspect. A direct check would compare the pipeline's pseudo-labels on synthetic target images against human annotations of the same images and require label accuracy at least as high as the final detector's AP50.

Watch

Extended reading notes

Core claim

The paper's central claim is that a fine-tuned latent diffusion model can serve as a data engine for cross-domain aerial vehicle detection: it can synthesize target-domain aerial images together with usable bounding-box labels, even though the model never sees target-box annotations. The label-generation mechanism is a chain of transfers through cross-attention maps. A detector is first trained on real source images with full boxes; it labels synthetic source images; a second detector is trained on the stacked cross-attention maps of those synthetic source images; and that second detector labels the stacked cross-attention maps of synthetic target images. The final, deployable detector is trained on synthetic target RGB images with those pseudo-labels and evaluated on real target images. The reported results support the claim that this label-transfer chain closes much of the domain gap.

Load-bearing premise

The load-bearing premise is that stacked cross-attention maps are nearly free of domain style, so a detector trained on synthetic source attention maps can correctly label synthetic target attention maps; the paper states this premise but runs no experiment that measures the transfer accuracy.

Editorial extensions

If this is right

  • On the two tested shifts, LINZ→UGRC and DOTA→UGRC, adding synthetic target images with attention-derived labels raises AP50 over source-only training by 4–23 percentage points across Faster R-CNN, YOLOv5, YOLOv8, and ViTDet.
  • The pipeline beats prior weakly supervised cross-domain detectors by 6–10 points and unsupervised domain adaptation by 7–40 points on the same shifts, using only image-level target labels.
  • Open-set detectors and large vision-language models perform poorly on aerial vehicles in their zero-shot settings, which positions synthetic data generation as a more practical adaptation route when target boxes are unavailable.
  • The label-refinement classifier makes the labeling robust to confidence-threshold differences across datasets, removing a threshold-tuning step that would otherwise be needed per target domain.

Reading between the lines

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

  • If the cross-attention maps are not as domain-invariant as assumed, the reported gains would mostly reflect having more training data and better synthetic images rather than successful label transfer; measuring source-vs-target attention-map shift would separate those explanations.
  • The pipeline should carry over to other small-object categories and other overhead imaging modalities, but the 112×112 pixel crop and 8×8 attention resolution set a floor on object size, so gains would shrink for extremely small or densely overlapping objects.
  • Because the evaluation metric is center-localization within a 12-pixel decision circle, the reported AP50 numbers may not reflect tight-box quality; users who care about precise footprint should request additional IoU thresholds.
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 / 4 minor

Summary. The paper addresses cross-domain vehicle detection in aerial imagery under a weak-supervision setting: fully labeled source images (LINZ or DOTA) and image-level target labels (UGRC). The proposed AGenDA pipeline fine-tunes a latent diffusion model on source and target images with domain-specific prompts, synthesizes 10k source and 10k target car images plus 10k empty target images, and labels the synthetic target images by a three-stage transfer: a detector FS trained on real source images pseudo-labels synthetic source images; a second detector FA is trained on stacked source cross-attention maps with those pseudo-labels; FA then predicts boxes on stacked target attention maps; the final detector is trained on synthetic target RGB images with these pseudo-labels and evaluated on the real target test set. Stacked attention maps combine the category word and two learnable tokens (foreground and background), with total-variation regularization. A classifier refinement module selects high/low-confidence labels. Experiments compare four detectors against source-only, open-set, UDA, and weakly supervised baselines, reporting AP50 gains (e.g., YOLOv8: 75.4% LINZ→UGRC and 75.7% DOTA→UGRC). The paper also introduces two new aerial datasets.

Significance. The proposed use of diffusion models for annotated aerial synthetic data is timely and the experimental scope is unusually broad: four detector backbones, two domain shifts, and multiple baseline families. The consistent double-digit gains over strong weak-supervision baselines (H2FA R-CNN, OCUD) are impressive if reproducible. However, the paper's central mechanism—the transfer of localization knowledge through grayscale cross-attention maps—is asserted rather than verified, and the evaluation rests on single-run AP50 numbers without uncertainty quantification. The new LINZ and UGRC datasets are a useful community resource, but they are not made available, and no code is released. The value of the contribution depends on closing these gaps.

major comments (4)
  1. [§5.2, Table 1] All AP50 values in Table 1 (and the ablations in Table 2 and Figure 5) are reported as single runs with no error bars, seeds, or significance tests. The reported advantage over the best weakly-supervised baseline is 6.2 and 7.2 percentage points, while several internal ablations differ by as little as 1.2 points (Table 2, DOTA→UGRC). Given stochasticity in detector and diffusion training, these differences are within the range where ranking could change across seeds. Please report mean±std over at least 3–5 seeds for the main comparisons and ablations, and state whether the headline improvements are statistically reliable.
  2. [§3.3, Algorithm 1] The core assumption of the label-transfer chain—that cross-attention maps are domain-invariant and localization-accurate enough to transfer a detector from source to target—is never directly measured. Step 5 assumes FS produces reliable pseudo-labels on synthetic source images; step 7 assumes FA trained on source attention maps transfers to target attention maps; and Section 3.3 justifies this by stating that attention maps 'contain less style information than RGB images,' but no experiment compares the source/target attention distributions or evaluates the resulting target pseudo-labels. Section 6 even concedes that the 8×8 attention resolution and overlapping objects make labeling difficult. Please add (i) quantitative label-quality evaluation of yGT against manually annotated boxes on synthetic target images, and (ii) an ablation that trains the final detector on synthetic target images labeled by an alternative mechanism (e.g., FS applied directly to synthetic target RGB, or attention thresholding without the FA transfer). Without these, the Table 1 gains cannot be attributed to the proposed attention-based transfer; they could be due to target-like synthetic RGB images with noisy labels.
  3. [§5.1 and §D.1] The paper replaces the stated decision-circle criterion (12 px radius) with a 42.36 px pseudo-bounding box and claims this is 'functionally equivalent' with 'minimal error,' but the error is never quantified. The isocontour IoU=0.5 region only matches the quarter-circle area, not its shape, so some centers inside the decision circle can be counted as false positives and vice versa. Because every number in Tables 1–3 and Figure 5 depends on this substitution, please report the worst-case or expected discrepancy between AP50 computed with the pseudo-box and AP computed directly with the decision circle, or include both metrics.
  4. [Datasets and reproducibility] The paper introduces two new datasets and a multi-stage pipeline but does not provide code, model weights, or a data release link (the project page URL is given, but no artifacts are described). Given the large number of implementation choices (Stage 1/2/3 hyperparameters in Table 4, thresholds, synthetic image counts), the central experimental claims are not independently reproducible without these artifacts. Please release code and the datasets (or a documented access procedure) at the project page.
minor comments (4)
  1. [Table 1] The row for YOLOv8 cites reference [15] (YOLOv5); it should cite [44]. Please check all table citations for consistency.
  2. [§3.3] The phrase 'use the well-trained detector FA to test the target domain cross-attention maps' should be 'to test on the target domain cross-attention maps'.
  3. [§5.1] The use of the phrase 'functionally equivalent' for the pseudo-bounding-box approximation is too strong without a quantified error bound; the wording should be softened or supported by the error analysis requested above.
  4. [Appendix D.2] The description says all training stages scale image resolution to 128×128 because YOLOv5 requires multiples of 32. Since the datasets are introduced at 112×112, please clarify whether evaluation is also performed at 128×128 and whether the 112×112 sampling size is preserved in any stage.

Circularity Check

1 steps flagged · score 2.0 of 10

Central AP50 results are not circular; one supplementary VLLM comparison uses a threshold optimized on the UGRC test set itself, a minor evaluation circularity.

  1. fitted input called prediction [Appendix C.3 (Table 3 comparison with VLLMs)]
    "To establish pseudo labels on the UGRC test set for our method, we set the detection threshold according to the highest F1 score achieved by each detector on UGRC test set."

    The threshold is selected by maximizing F1 on the UGRC test labels that are then used to compute the reported precision/recall for 'Ours' in Table 3. Consequently, those precision/recall values (e.g., YOLOv8 70.0/76.3) are not an independent evaluation of the trained detector; they are the result of an oracle-threshold optimization on the test set itself. The VLLM comparison in Table 3 is therefore forced by construction for these metrics. This does not affect the central AP50 claims in Table 1, because AP averages over confidence thresholds and the labeling thresholds in Sec. 5.1 were instead set on synthetic source cross-attention F1.

full rationale

The main derivation is not circular. The pseudo-label chain in Algorithm 1 flows from source bounding boxes to synthetic source images, then to source attention maps, then to target attention maps, then to synthetic target images, and is finally evaluated on real target images with manual labels; no equation defines the final AP50 in terms of the inputs. The confidence thresholds used for automatic labeling are chosen from synthetic source cross-attention F1 (Sec. 5.1), not from target labels. The Sec. 3.3 claim that attention maps are 'style-less' is an unmeasured empirical premise, but it is an assumption about transferability, not a logical loop. The paper's own Sec. 6 limitations (8x8 attention resolution, overlapping objects) are acknowledged and independent of circularity. The only concrete circular step is in Appendix C.3, where the threshold for 'Ours' in the VLLM precision/recall comparison is set by maximizing F1 on the UGRC test set itself, making those specific metrics fitted rather than predicted; the Table 1 AP50 conclusions remain externally evaluated and non-circular.

Assumptions & free parameters 4 free parameters · 5 assumptions · 0 invented entities

The central claim rests on learned prompt tokens and thresholds (free parameters), on the assumption that attention maps localize and transfer across domains, and on a metric approximation. No new physical entities are introduced.

free parameters (4)
  • Learnable prompt tokens [V1], [V2], [V3] = Not reported numerically (learned embeddings)
    Trained in stage 1 to encode foreground car and background concepts; central to multi-channel cross-attention label quality (Section 3.2).
  • Label refinement thresholds lambda_high and lambda_low = 0.7/0.3 for YOLOv5 and YOLOv8; 0.95/0.5 for Faster-RCNN and ViTDet
    Chosen as the threshold yielding the best F1 score on synthetic source cross-attention maps during detector training (Section 5.1).
  • Number of synthetic images per domain = 10,000 with cars for source and target; 10,000 without cars for target
    Hand-chosen dataset sizes for generated training data (Section 5.1).
  • Pseudo-bounding box size = 42.36 px
    Derived to approximate the 12 px radius decision circle under IoU=0.5, but the approximation error is not quantified (Section 5.1, Appendix D.1).
assumptions (5)
  • domain assumption Cross-attention maps in a fine-tuned text-to-image diffusion model highlight regions corresponding to the text concept.
    Adopted from prior work [35,54] and assumed to hold for 112px aerial patches (Section 3.2).
  • domain assumption Cross-attention maps are style-less enough that a detector trained on source attention maps transfers to target attention maps.
    Stated in Section 3.3 without direct measurement; the whole label-transfer chain depends on it.
  • domain assumption Fine-tuned Stable Diffusion generates synthetic target images that share a distribution with real target images sufficiently for detector training.
    The pipeline relies on synthetic target images acting as a proxy for real target data (Section 3.1).
  • domain assumption Small vehicles are representable by the 8x8 cross-attention grid after the 112px patch design.
    The conclusion notes that target objects may occupy less than a 1x1 grid cell, so the positive result depends on this assumption holding.
  • ad hoc to paper The 42.36 px pseudo-bounding box with IoU>=0.5 is functionally equivalent to the 12 px decision circle for AP50.
    Appendix D.1 matches the area under the IoU=0.5 isocontour to one quarter of the decision circle; the claimed minimal error is asserted, not bounded.

how reviews work

0 comments
Cite this review

Pith. "Pith review of Adapting Vehicle Detectors for Aerial Imagery to Unseen Domains with Weak Supervision." pith.science (2026). https://pith.science/paper/XWVDJ62Z

@misc{pith2026250720976,
  author       = {Pith},
  title        = {Pith review of: Adapting Vehicle Detectors for Aerial Imagery to Unseen Domains with Weak Supervision},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/XWVDJ62Z}},
  note         = {Machine review of arXiv:2507.20976}
}
read the original abstract

Detecting vehicles in aerial imagery is a critical task with applications in traffic monitoring, urban planning, and defense intelligence. Deep learning methods have provided state-of-the-art (SOTA) results for this application. However, a significant challenge arises when models trained on data from one geographic region fail to generalize effectively to other areas. Variability in factors such as environmental conditions, urban layouts, road networks, vehicle types, and image acquisition parameters (e.g., resolution, lighting, and angle) leads to domain shifts that degrade model performance. This paper proposes a novel method that uses generative AI to synthesize high-quality aerial images and their labels, improving detector training through data augmentation. Our key contribution is the development of a multi-stage, multi-modal knowledge transfer framework utilizing fine-tuned latent diffusion models (LDMs) to mitigate the distribution gap between the source and target environments. Extensive experiments across diverse aerial imagery domains show consistent performance improvements in AP50 over supervised learning on source domain data, weakly supervised adaptation methods, unsupervised domain adaptation methods, and open-set object detectors by 4-23%, 6-10%, 7-40%, and more than 50%, respectively. Furthermore, we introduce two newly annotated aerial datasets from New Zealand and Utah to support further research in this field. Project page is available at: https://humansensinglab.github.io/AGenDA

Figures

Figures reproduced from arXiv: 2507.20976 by the authors.

Figure 1
Figure 1. Overview. We propose a pipeline that generates high￾quality aerial images along with their labels. Our method outper￾forms baseline detectors trained on source images and open-set detectors directly inferred on target images. nearly indistinguishable from real-world data [12, 36, 52]. Additionally, integrating natural human language has be￾come crucial to modern vision systems. This integration can take the form of … view at source ↗
Figure 2
Figure 2. Overview of our pipeline. It consists of two stages. First, we finetune Stable Diffusion and synthesize both source and target domain images. Second, we automatically label synthetic target domain images via cross-attention maps. represents the object type, and [S] and [T] are unique iden￾tifiers to distinguish between the source and target domain prompts, as inspired by [47]. The U-Net ϵθ is trained to predict the … view at source ↗
Figure 3
Figure 3. Image samples from our datasets. (left) LINZ sample, (right) UGRC, (green markers) small vehicle location annotations. For more examples, check the Supplementary Material. respectively. The original bounding box labels were con￾verted to object locations to match the annotations of the other two datasets described below. LINZ: We created this dataset by manually annotating ve￾hicle locations in aerial images capture… view at source ↗
Figures from the paper (12 more)
Figure 4
Figure 4. Figure 4: Cross-attention maps for different tokens. We analyze the effectiveness of utilizing multi-channel cross-attention maps by assessing the label quality synthetic UGRC images. (a) Synthetic UGRC images. (b) Labels generated using only the cross-attention map of the word …
Figure 5
Figure 5. Figure 5: Quantitative comparison with varying thresholds. We report the AP50 result. (a) (b) (c) (d) (e) [PITH_FULL_IMAGE:figures/full_fig_p008_5.png]
Figure 6
Figure 6. Figure 6: Comparison of our method with different thresholds for label quality. Blue bounding boxes represent pseudo labels generated using fixed thresholds, while red bounding boxes cor￾respond to labels obtained through refinement. The blue and red dots indicate the predicted …
Figure 7
Figure 7. Figure 7: Vehicles belonging to the object class small vehicle. amples of both LINZ and UGRC images, highlighting dis￾tinct visual characteristics: UGRC includes a notable pro￾portion of off-road vehicles, reflecting its sandy and rocky terrain, while LINZ images primarily featu…
Figure 8
Figure 8. Figure 8: Geographic regions where we construct LINZ and UGRC datasets. Red bounding boxes denote the testing area. [PITH_FULL_IMAGE:figures/full_fig_p014_8.png]
Figure 9
Figure 9. Figure 9: Examples of images from our real-world datasets. (first row) LINZ images containing small vehicles; (second row) LINZ images without vehicles; (third row) UGRC images containing small vehicles; (fourth row) UGRC images without vehicles; LINZ UGRC 63.2 88.9 50 60 70 80 …
Figure 10
Figure 10. Figure 10: Comparison between cross-dataset generalization and within-dataset performance. The purple bars represent the model trained on the LINZ dataset and evaluated on the UGRC dataset, while the pink bars correspond to both training and testing conducted on the UGRC dataset…
Figure 12
Figure 12. Figure 12: Two popular VLLMs (BLIP2 and Kosmos2) tested as [PITH_FULL_IMAGE:figures/full_fig_p016_12.png]
Figure 11
Figure 11. Figure 11: VLMs captioning capabilities analysis tested on LINZ [PITH_FULL_IMAGE:figures/full_fig_p016_11.png]
Figure 13
Figure 13. Figure 13: Failure cases of Open-set detectors. (a) Detection results of Grounding-DINO. (b) Detection results of Omdet-Turbo. (c) [PITH_FULL_IMAGE:figures/full_fig_p017_13.png]
Figure 14
Figure 14. Figure 14: Illustration of how we obtain the 42.36 px bounding box size. (a) The black bounding boxes denote the ground truth pseudo [PITH_FULL_IMAGE:figures/full_fig_p017_14.png]
Figure 15
Figure 15. Figure 15: Failure cases of diffusion models. (a) Images generated by pre-trained Stable Diffusion V1.4. (b) Images generated by pre [PITH_FULL_IMAGE:figures/full_fig_p018_15.png]

Discussion (0). Continue with ORCID to comment.

Reference graph

Works this paper leans on

75 extracted references · 56 canonical work pages

  1. [1]

    Qwen2.5-VL technical report

    Shuai Bai, Keqin Chen, Xuejing Liu, Jialin Wang, Wenbin Ge, Sibo Song, Kai Dang, Peng Wang, Shijie Wang, Jun Tang, et al. Qwen2.5-VL technical report. arXiv preprint arXiv:2502.13923, 2025. 2

    arXiv 2025

  2. [2]

    MMDetection: Open mmlab detection toolbox and benchmark

    Kai Chen, Jiaqi Wang, Jiangmiao Pang, Yuhang Cao, Yu Xiong, Xiaoxiao Li, Shuyang Sun, Wansen Feng, Ziwei Liu, Jiarui Xu, Zheng Zhang, Dazhi Cheng, Chenchen Zhu, Tian- heng Cheng, Qijie Zhao, Buyu Li, Xin Lu, Rui Zhu, Yue Wu, Jifeng Dai, Jingdong Wang, Jianping Shi, Wanli Ouyang, Chen Change Loy, and Dahua Lin. MMDetection: Open mmlab detection toolbox and...

    arXiv 1906

  3. [3]

    GeoDiffusion: Text-Prompted Geometric Control for Object Detection Data Generation

    Kai Chen, Enze Xie, Zhe Chen, Yibo Wang, Lanqing HONG, Zhenguo Li, and Dit-Yan Yeung. GeoDiffusion: Text-Prompted Geometric Control for Object Detection Data Generation. In The Twelfth International Conference on Learning Representations, 2024. 2

    work page 2024

  4. [4]

    YOLO-World: Real-Time Open-V ocabulary Object Detection

    Tianheng Cheng, Lin Song, Yixiao Ge, Wenyu Liu, Xing- gang Wang, and Ying Shan. YOLO-World: Real-Time Open-V ocabulary Object Detection. In Proc. IEEE Conf. Computer Vision and Pattern Recognition (CVPR), 2024. 3

    work page 2024

  5. [5]

    Gemini 2.5: Pushing the frontier with advanced reasoning, multimodality, long context, and next generation agentic capabilities

    Gheorghe Comanici, Eric Bieber, Mike Schaekermann, Ice Pasupat, Noveen Sachdeva, Inderjit Dhillon, Marcel Blis- tein, Ori Ram, Dan Zhang, Evan Rosen, et al. Gemini 2.5: Pushing the frontier with advanced reasoning, multimodality, long context, and next generation agentic capabilities. arXiv preprint arXiv:2507.06261, 2025. 2

    arXiv 2025

  6. [6]

    Diffusion Models Beat GANs on Image Synthesis

    Prafulla Dhariwal and Alexander Quinn Nichol. Diffusion Models Beat GANs on Image Synthesis. In Advances in Neural Information Processing Systems, 2021. 2

    work page 2021

  7. [7]

    LaMI-DETR: Open-V ocabulary Detection with Lan- guage Model Instruction

    Penghui Du, Yu Wang, Yifan Sun, Luting Wang, Yue Liao, Gang Zhang, Errui Ding, Yan Wang, Jingdong Wang, and Si Liu. LaMI-DETR: Open-V ocabulary Detection with Lan- guage Model Instruction. In Proceedings of the European conference on computer vision (ECCV), 2024. 3

    work page 2024

  8. [8]

    Diversify your vision datasets with automatic diffusion-based augmentation

    Lisa Dunlap, Alyssa Umino, Han Zhang, Jiezhi Yang, Joseph E Gonzalez, and Trevor Darrell. Diversify your vision datasets with automatic diffusion-based augmentation. In Advances in Neural Information Processing Systems , pages 79024–79034. Curran Associates, Inc., 2023. 2

    work page 2023

Show all 75 references
  1. [9]

    Everingham, L

    M. Everingham, L. Van Gool, C. K. I. Williams, J. Winn, and A. Zisserman. The PASCAL Visual Object Classes Challenge 2012 (VOC2012) Results. http://www.pascal- network.org/challenges/VOC/voc2012/workshop/index.html,

    2012

  2. [10]

    An Image is Worth One Word: Personalizing Text-to-Image Generation using Textual Inversion

    Rinon Gal, Yuval Alaluf, Yuval Atzmon, Or Patashnik, Amit Haim Bermano, Gal Chechik, and Daniel Cohen-or. An Image is Worth One Word: Personalizing Text-to-Image Generation using Textual Inversion. In The Eleventh Inter- national Conference on Learning Representations, 2023. 4

    2023

  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 (CVPR), 2016. 6

    2016

  4. [12]

    Denoising Dif- fusion Probabilistic Models

    Jonathan Ho, Ajay Jain, and Pieter Abbeel. Denoising Dif- fusion Probabilistic Models. In Advances in Neural Infor- mation Processing Systems, pages 6840–6851. Curran Asso- ciates, Inc., 2020. 1, 2, 3

    2020

  5. [13]

    Cross-domain weakly-supervised object de- tection through progressive domain adaptation

    Naoto Inoue, Ryosuke Furuta, Toshihiko Yamasaki, and Kiy- oharu Aizawa. Cross-domain weakly-supervised object de- tection through progressive domain adaptation. In Proceed- ings of the IEEE conference on computer vision and pattern recognition, pages 5001–5009, 2018. 3

    2018

  6. [14]

    DGIn- Style: Domain-Generalizable Semantic Segmentation with Image Diffusion Models and Stylized Semantic Control

    Yuru Jia, Lukas Hoyer, Shengyu Huang, Tianfu Wang, Luc Van Gool, Konrad Schindler, and Anton Obukhov. DGIn- Style: Domain-Generalizable Semantic Segmentation with Image Diffusion Models and Stylized Semantic Control. In Computer Vision – ECCV 2024, pages 91–109, Cham, 2025. Sp...

    2024

  7. [15]

    Yolov5 by ultralytics, 2020

    Glenn Jocher. Yolov5 by ultralytics, 2020. 6, 7, 2

    2020

  8. [16]

    Align and distill: Unifying and improving domain adaptive object detection

    Justin Kay, Timm Haucke, Suzanne Stathatos, Siqi Deng, Erik Young, Pietro Perona, Sara Beery, and Grant Van Horn. Align and distill: Unifying and improving domain adaptive object detection. arXiv preprint arXiv:2403.12029, 2024. 3

    2024 arXiv

  9. [17]

    Lobell, and Ste- fano Ermon

    Samar Khanna, Patrick Liu, Linqi Zhou, Chenlin Meng, Robin Rombach, Marshall Burke, David B. Lobell, and Ste- fano Ermon. DiffusionSat: A Generative Foundation Model for Satellite Imagery. In The Twelfth International Confer- ence on Learning Representations, 2024. 2

    2024

  10. [18]

    Text-Image Alignment for Diffusion-Based Perception

    Neehar Kondapaneni, Markus Marks, Manuel Knott, Roge- rio Guimaraes, and Pietro Perona. Text-Image Alignment for Diffusion-Based Perception. Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR), pages 13883–13893, 2024. 3

    2024

  11. [19]

    Scaling novel object detection with weakly su- pervised detection transformers

    Tyler LaBonte, Yale Song, Xin Wang, Vibhav Vineet, and Neel Joshi. Scaling novel object detection with weakly su- pervised detection transformers. In 2023 IEEE/CVF Win- ter Conference on Applications of Computer Vision (WACV), pages 85–96, 2023. 3

    2023

  12. [20]

    Markov chains and mixing times

    David A Levin and Yuval Peres. Markov chains and mixing times. American Mathematical Soc., 2017. 5

    2017

  13. [21]

    Your diffusion model is secretly a zero-shot classifier

    Alexander C Li, Mihir Prabhudesai, Shivam Duggal, Ellis Brown, and Deepak Pathak. Your diffusion model is secretly a zero-shot classifier. In Proceedings of the IEEE/CVF Inter- national Conference on Computer Vision, pages 2206–2217,

  14. [22]

    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. In ICML,

  15. [23]

    Grounded language-image pre-training

    Liunian Harold Li*, Pengchuan Zhang*, Haotian Zhang*, Jianwei Yang, Chunyuan Li, Yiwu Zhong, Lijuan Wang, Lu Yuan, Lei Zhang, Jenq-Neng Hwang, Kai-Wei Chang, and Jianfeng Gao. Grounded language-image pre-training. In CVPR, 2022. 2, 3, 7

    2022

  16. [24]

    Sigma: Semantic- complete graph matching for domain adaptive object detec- tion

    Wuyang Li, Xinyu Liu, and Yixuan Yuan. Sigma: Semantic- complete graph matching for domain adaptive object detec- tion. In CVPR, 2022. 3, 7

    2022

  17. [25]

    Exploring plain vision transformer backbones for object de- tection

    Yanghao Li, Hanzi Mao, Ross Girshick, and Kaiming He. Exploring plain vision transformer backbones for object de- tection. In Computer Vision – ECCV 2022, pages 280–296, Cham, 2022. Springer Nature Switzerland. 6, 7, 2

    2022

  18. [26]

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

    Yuheng Li, Haotian Liu, Qingyang Wu, Fangzhou Mu, Jian- wei Yang, Jianfeng Gao, Chunyuan Li, and Yong Jae Lee. Gligen: Open-set grounded text-to-image generation. InPro- ceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR), pages 22511–22521, 2...

    2023

  19. [27]

    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 Va- jda. Cross-Domain Adaptive Teacher for Object Detection. In IEEE Conference on Computer Vision and Pattern Recog- nition (CVPR), 2022. 3, 7

    2022

  20. [28]

    Microsoft coco: Common objects in context

    Tsung-Yi Lin, Michael Maire, Serge Belongie, James Hays, Pietro Perona, Deva Ramanan, Piotr Doll´ar, and C Lawrence Zitnick. Microsoft coco: Common objects in context. In Computer vision–ECCV 2014: 13th European conference, zurich, Switzerland, September 6-12, 2014, proceeding...

    2014

  21. [29]

    Selwyn 0.125m Urban Aerial Photos (2012-2013)

    LINZ. Selwyn 0.125m Urban Aerial Photos (2012-2013). https://data.linz.govt.nz/layer/51926- selwyn- 0125m- urban- aerial- photos- 2012- 2013/, 2014. 6

    2012

  22. [30]

    LLaV A-NeXT: Im- proved reasoning, OCR, and world knowledge, 2024

    Haotian Liu, Chunyuan Li, Yuheng Li, Bo Li, Yuanhan Zhang, Sheng Shen, and Yong Jae Lee. LLaV A-NeXT: Im- proved reasoning, OCR, and world knowledge, 2024. 2

    2024

  23. [31]

    Grounding dino: Marrying DINO with Grounded Pre-training for Open-Set Object Detection

    Shilong Liu, Zhaoyang Zeng, Tianhe Ren, Feng Li, Hao Zhang, Jie Yang, Qing Jiang, Chunyuan Li, Jianwei Yang, Hang Su, et al. Grounding dino: Marrying DINO with Grounded Pre-training for Open-Set Object Detection. In European Conference on Computer Vision , pages 38–55. Springe...

    2024

  24. [32]

    Swin transformer: Hierarchical vision transformer using shifted windows

    Ze Liu, Yutong Lin, Yue Cao, Han Hu, Yixuan Wei, Zheng Zhang, Stephen Lin, and Baining Guo. Swin transformer: Hierarchical vision transformer using shifted windows. In Proceedings of the IEEE/CVF International Conference on Computer Vision (ICCV), 2021. 7

    2021

  25. [33]

    Simple open-vocabulary object detection

    Matthias Minderer, Alexey Gritsenko, Austin Stone, Maxim Neumann, Dirk Weissenborn, Alexey Dosovitskiy, Aravindh Mahendran, Anurag Arnab, Mostafa Dehghani, Zhuoran Shen, Xiao Wang, Xiaohua Zhai, Thomas Kipf, and Neil Houlsby. Simple open-vocabulary object detection. In Com- pu...

    2022

  26. [34]

    Scaling Open-V ocabulary Object Detection

    Matthias Minderer, Alexey Gritsenko, and Neil Houlsby. Scaling Open-V ocabulary Object Detection. In Advances in Neural Information Processing Systems , pages 72983– 73007. Curran Associates, Inc., 2023. 2, 3, 7

    2023

  27. [35]

    Dataset Diffusion: Diffusion-based Synthetic Data Generation for Pixel-Level Semantic Segmentation

    Quang Ho Nguyen, Truong Tuan Vu, Anh Tuan Tran, and Khoi Nguyen. Dataset Diffusion: Diffusion-based Synthetic Data Generation for Pixel-Level Semantic Segmentation. In Thirty-seventh Conference on Neural Information Process- ing Systems, 2023. 2, 3

    2023

  28. [36]

    Improved denoising diffusion probabilistic models

    Alexander Quinn Nichol and Prafulla Dhariwal. Improved denoising diffusion probabilistic models. In International conference on machine learning, pages 8162–8171. PMLR,

  29. [37]

    AttnDreamBooth: To- wards Text-Aligned Personalized Text-to-Image Generation

    Lianyu Pang, Jian Yin, Baoquan Zhao, Feize Wu, Fu Lee Wang, Qing Li, and Xudong Mao. AttnDreamBooth: To- wards Text-Aligned Personalized Text-to-Image Generation. In The Thirty-eighth Annual Conference on Neural Informa- tion Processing Systems, 2024. 3, 4

    2024

  30. [38]

    Shape-Guided Diffusion With Inside-Outside Atten- tion

    Dong Huk Park, Grace Luo, Clayton Toste, Samaneh Azadi, Xihui Liu, Maka Karalashvili, Anna Rohrbach, and Trevor Darrell. Shape-Guided Diffusion With Inside-Outside Atten- tion. In Proceedings of the IEEE/CVF Winter Conference on Applications of Computer Vision (WACV), pages 4198–4207,

  31. [39]

    One-step image translation with text-to-image models

    Gaurav Parmar, Taesung Park, Srinivasa Narasimhan, and Jun-Yan Zhu. One-step image translation with text-to-image models. arXiv preprint arXiv:2403.12036, 2024. 3, 7

    2024 arXiv

  32. [40]

    Ground- ing multimodal large language models to the world

    Zhiliang Peng, Wenhui Wang, Li Dong, Yaru Hao, Shaohan Huang, Shuming Ma, Qixiang Ye, and Furu Wei. Ground- ing multimodal large language models to the world. In The Twelfth International Conference on Learning Representa- tions, 2024. 3

    2024

  33. [41]

    SDXL: Improving latent diffusion models for high-resolution image synthesis

    Dustin Podell, Zion English, Kyle Lacey, Andreas Blattmann, Tim Dockhorn, Jonas M ¨uller, Joe Penna, and Robin Rombach. SDXL: Improving latent diffusion models for high-resolution image synthesis. In The Twelfth Inter- national Conference on Learning Representations, 2024. 1, 2

    2024

  34. [42]

    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

  35. [43]

    Hierarchical text-conditional image gener- ation with clip latents

    Aditya Ramesh, Prafulla Dhariwal, Alex Nichol, Casey Chu, and Mark Chen. Hierarchical text-conditional image gener- ation with clip latents. arXiv preprint arXiv:2204.06125, 1 (2):3, 2022. 2, 6

    2022 arXiv

  36. [44]

    Real-time flying object detection with yolov8.arXiv preprint arXiv:2305.09972, 2023

    Dillon Reis, Jordan Kupec, Jacqueline Hong, and Ahmad Daoudi. Real-time flying object detection with yolov8.arXiv preprint arXiv:2305.09972, 2023. 6, 7, 2

    2023 arXiv

  37. [45]

    Faster R-CNN: Towards Real-Time Object Detection with Region Proposal Networks

    Shaoqing Ren, Kaiming He, Ross Girshick, and Jian Sun. Faster R-CNN: Towards Real-Time Object Detection with Region Proposal Networks. In Advances in Neural Informa- tion Processing Systems. Curran Associates, Inc., 2015. 6, 7, 2

    2015

  38. [46]

    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), pages 10684–10695, 2022. 1, 2, 3, 6, 7

    2022

  39. [47]

    DreamBooth: Fine Tuning Text-to-Image Diffusion Models for Subject-Driven Generation

    Nataniel Ruiz, Yuanzhen Li, Varun Jampani, Yael Pritch, Michael Rubinstein, and Kfir Aberman. DreamBooth: Fine Tuning Text-to-Image Diffusion Models for Subject-Driven Generation. In Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR), page...

    2023

  40. [48]

    Imagenet large scale visual recognition challenge

    Olga Russakovsky, Jia Deng, Hao Su, Jonathan Krause, San- jeev Satheesh, Sean Ma, Zhiheng Huang, Andrej Karpathy, Aditya Khosla, Michael Bernstein, et al. Imagenet large scale visual recognition challenge. International journal of computer vision, 115:211–252, 2015. 6

    2015

  41. [49]

    LAION-5b: An open large-scale dataset for train- ing next generation image-text models

    Christoph Schuhmann, Romain Beaumont, Richard Vencu, Cade W Gordon, Ross Wightman, Mehdi Cherti, Theo Coombes, Aarush Katta, Clayton Mullis, Mitchell Worts- man, Patrick Schramowski, Srivatsa R Kundurthy, Katherine Crowson, Ludwig Schmidt, Robert Kaczmarczyk, and Jenia Jitsev....

    2022

  42. [50]

    Deep unsupervised learning using nonequilibrium thermodynamics

    Jascha Sohl-Dickstein, Eric Weiss, Niru Maheswaranathan, and Surya Ganguli. Deep unsupervised learning using nonequilibrium thermodynamics. In International confer- ence on machine learning, pages 2256–2265. pmlr, 2015. 1

    2015

  43. [51]

    Satdiffmoe: A mixture of estimation method for satellite image super- resolution with latent diffusion models

    Bowen Song, Zhaoxu Luo, and Liyue Shen. Satdiffmoe: A mixture of estimation method for satellite image super- resolution with latent diffusion models. InICML 2024 Work- shop on Structured Probabilistic Inference & Generative Modeling, 2024. 2

    2024

  44. [52]

    Denois- ing diffusion implicit models

    Jiaming Song, Chenlin Meng, and Stefano Ermon. Denois- ing diffusion implicit models. In International Conference on Learning Representations, 2021. 1

    2021

  45. [53]

    Multiple Instance Detection Network With Online Instance Classifier Refinement

    Peng Tang, Xinggang Wang, Xiang Bai, and Wenyu Liu. Multiple Instance Detection Network With Online Instance Classifier Refinement. In Proceedings of the IEEE Confer- ence on Computer Vision and Pattern Recognition (CVPR) ,

  46. [54]

    What the DAAM: Interpreting Stable Dif- fusion Using Cross Attention

    Raphael Tang, Linqing Liu, Akshat Pandey, Zhiying Jiang, Gefei Yang, Karun Kumar, Pontus Stenetorp, Jimmy Lin, and Ferhan Ture. What the DAAM: Interpreting Stable Dif- fusion Using Cross Attention. In Proceedings of the 61st An- nual Meeting of the Association for Computationa...

    2023

  47. [55]

    Gemini: a family of highly capable multimodal models

    Gemini Team, Rohan Anil, Sebastian Borgeaud, Jean- Baptiste Alayrac, Jiahui Yu, Radu Soricut, Johan Schalkwyk, Andrew M Dai, Anja Hauth, Katie Millican, et al. Gemini: a family of highly capable multimodal models. arXiv preprint arXiv:2312.11805, 2023. 2

    2023 arXiv

  48. [56]

    Gemini 1.5: Unlocking multimodal understanding across millions of tokens of con- text

    Gemini Team, Petko Georgiev, Ving Ian Lei, Ryan Burnell, Libin Bai, Anmol Gulati, Garrett Tanzer, Damien Vincent, Zhufeng Pan, Shibo Wang, et al. Gemini 1.5: Unlocking multimodal understanding across millions of tokens of con- text. arXiv preprint arXiv:2403.05530, 2024. 2

    2024 arXiv

  49. [57]

    Utah High Resolution Orthophotography (HRO) 2012 Images

    UGRC. Utah High Resolution Orthophotography (HRO) 2012 Images. https : / / gis . utah . gov / products / sgid / aerial - photography / high - resolution-orthophotography/, 2012. 6

    2012

  50. [58]

    Diffusion model is secretly a training-free open vocabulary semantic segmenter

    Jinglong Wang, Xiawei Li, Jing Zhang, Qingyuan Xu, Qin Zhou, Qian Yu, Lu Sheng, and Dong Xu. Diffusion model is secretly a training-free open vocabulary semantic segmenter. CoRR, abs/2309.02773, 2023. 3

    2023 arXiv

  51. [59]

    Domain Gap Embeddings for Genera- tive Dataset Augmentation

    Yinong Oliver Wang, Younjoon Chung, Chen Henry Wu, and Fernando De la Torre. Domain Gap Embeddings for Genera- tive Dataset Augmentation. In Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR), pages 28684–28694, 2024. 2, 3

    2024

  52. [60]

    DatasetDM: Synthesizing Data with Perception An- notations Using Diffusion Models

    Weijia Wu, Yuzhong Zhao, Hao Chen, Yuchao Gu, Rui Zhao, Yefei He, Hong Zhou, Mike Zheng Shou, and Chunhua Shen. DatasetDM: Synthesizing Data with Perception An- notations Using Diffusion Models. In Thirty-seventh Con- ference on Neural Information Processing Systems, 2023. 2, 3

    2023

  53. [61]

    DiffuMask: Synthesizing Images with Pixel-level Annotations for Semantic Segmentation Using Diffusion Models

    Weijia Wu, Yuzhong Zhao, Mike Zheng Shou, Hong Zhou, and Chunhua Shen. DiffuMask: Synthesizing Images with Pixel-level Annotations for Semantic Segmentation Using Diffusion Models. In Proceedings of the IEEE/CVF Interna- tional Conference on Computer Vision (ICCV), pages 1206–...

    2023

  54. [62]

    SOEDiff: Efficient Distillation for Small Object Editing

    Yiming Wu, Qihe Pan, Zhen Zhao, Zicheng Wang, Sifan Long, and Ronghua Liang. SOEDiff: Efficient Distillation for Small Object Editing. ACM Trans. Multimedia Comput. Commun. Appl., 2025. 2

    2025

  55. [63]

    Deepseek-vl2: Mixture-of- experts vision-language models for advanced multimodal understanding

    Zhiyu Wu, Xiaokang Chen, Zizheng Pan, Xingchao Liu, Wen Liu, Damai Dai, Huazuo Gao, Yiyang Ma, Chengyue Wu, Bingxuan Wang, et al. Deepseek-vl2: Mixture-of- experts vision-language models for advanced multimodal understanding. arXiv preprint arXiv:2412.10302, 2024. 2

    2024 arXiv

  56. [64]

    DOTA: A Large-Scale Dataset for Object Detec- tion in Aerial Images

    Gui-Song Xia, Xiang Bai, Jian Ding, Zhen Zhu, Serge Be- longie, Jiebo Luo, Mihai Datcu, Marcello Pelillo, and Liang- pei Zhang. DOTA: A Large-Scale Dataset for Object Detec- tion in Aerial Images. In Proceedings of the IEEE Confer- ence on Computer Vision and Pattern Recogniti...

  57. [65]

    CycleNet: Rethinking Cycle Consistency in Text-Guided Diffusion for Image Manipulation

    Sihan Xu, Ziqiao Ma, Yidong Huang, Honglak Lee, and Joyce Chai. CycleNet: Rethinking Cycle Consistency in Text-Guided Diffusion for Image Manipulation. In Thirty- seventh Conference on Neural Information Processing Sys- tems, 2023. 3

    2023

  58. [66]

    H2FA R-CNN: Holistic and Hierarchical Feature Alignment for Cross-Domain Weakly Supervised Object De- tection

    Yunqiu Xu, Yifan Sun, Zongxin Yang, Jiaxu Miao, and Yi Yang. H2FA R-CNN: Holistic and Hierarchical Feature Alignment for Cross-Domain Weakly Supervised Object De- tection. In Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR) , pages 14329...

    2022

  59. [67]

    A survey on multimodal large language models

    Shukang Yin, Chaoyou Fu, Sirui Zhao, Ke Li, Xing Sun, Tong Xu, and Enhong Chen. A survey on multimodal large language models. National Science Review, 11(12), 2024. 2

    2024

  60. [68]

    Sigmoid Loss for Language Image Pre- training

    Xiaohua Zhai, Basil Mustafa, Alexander Kolesnikov, and Lucas Beyer. Sigmoid Loss for Language Image Pre- training. In Proceedings of the IEEE/CVF international con- ference on computer vision, pages 11975–11986, 2023. 2

    2023

  61. [69]

    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 (ICCV), pages 3836–3847, 2023. 2

    2023

  62. [70]

    Diffusionengine: Diffusion model is scalable data engine for object detection

    Manlin Zhang, Jie Wu, Yuxi Ren, Ming Li, Jie Qin, Xue- feng Xiao, Wei Liu, Rui Wang, Min Zheng, and Andy J Ma. Diffusionengine: Diffusion model is scalable data engine for object detection. arXiv preprint arXiv:2309.03893, 2023. 2

    2023 arXiv

  63. [71]

    Task-Specific Inconsistency Alignment for Domain Adaptive Object Detection

    Liang Zhao and Limin Wang. Task-Specific Inconsistency Alignment for Domain Adaptive Object Detection. In Pro- ceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR), pages 14217–14226, 2022. 3, 7

    2022

  64. [72]

    Real-time Transformer-based Open-V ocabulary Detection with Efficient Fusion Head

    Tiancheng Zhao, Peng Liu, Xuan He, Lu Zhang, and Kyu- song Lee. Real-time Transformer-based Open-V ocabulary Detection with Efficient Fusion Head. arXiv preprint arXiv:2403.06892, 2024. 2, 3, 7

    2024 arXiv

  65. [73]

    Boosting weakly supervised object detection with progres- sive knowledge transfer

    Yuanyi Zhong, Jianfeng Wang, Jian Peng, and Lei Zhang. Boosting weakly supervised object detection with progres- sive knowledge transfer. In Computer Vision – ECCV 2020, pages 615–631, Cham, 2020. Springer International Publish- ing. 7

    2020

  66. [74]

    SSDA-YOLO: Semi-supervised domain adaptive YOLO for cross-domain object detection

    Huayi Zhou, Fei Jiang, and Hongtao Lu. SSDA-YOLO: Semi-supervised domain adaptive YOLO for cross-domain object detection. Computer Vision and Image Understand- ing, 229:103649, 2023. 3, 7

    2023

  67. [75]

    an aerial image with cars in Utah

    Jinguo Zhu, Weiyun Wang, Zhe Chen, Zhaoyang Liu, Shen- glong Ye, Lixin Gu, Hao Tian, Yuchen Duan, Weijie Su, Jie Shao, et al. InternVL3: Exploring advanced training and test-time recipes for open-source multimodal models. arXiv preprint arXiv:2504.10479, 2025. 2 Adapting Vehic...

    2025 arXiv

Pith tools

Reviewed August 6, 2026 · model on record in the stance chip above.