Pith. sign in

REVIEW 5 major objections 6 minor 38 references

The paper claims that a foundation-model-collaborative active learning loop, built on a dual-source mechanism (cached UPN+SAM2 reference boxes plus current detector predictions), achieves superior or comparable detection performance under m

Reviewed by Pith at T0; open to challenge. T0 means a machine referee read the full paper against a public rubric. the ladder, T0–T4 →

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

A foundation-model-assisted active learning system using UPN+SAM2 reference boxes and DINOv2 object features improves low-budget annotation efficiency for remote sensing object detection.

T0 review reviewed 2026-08-01 challenge →

load-bearing objection Serious engineering with a load-bearing but under-validated SA-source; referee it, but require code, a citation fix, and a sensitivity analysis. the 5 major comments →

arxiv 2607.16671 v1 pith:N56HEYHD submitted 2026-07-18 cs.CV

Foundation-Assisted Active Learning for Object Detection Annotation

Jinchang Zhang , Arnold Zumbrun , Jing Lin , Guoyu Lu This is my paper
classification cs.CV
keywords active learningobject detectionremote sensingfoundation modelsSAM2UPNsemi-automatic annotationsample efficiency
verification ladder T0 review T1 audit T2 compute T3 formal T4 reserved

The pith

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

The reading

This paper argues that the main obstacle to efficient remote sensing object detection annotation is not classification uncertainty but cold-start localization noise: early detectors classify reasonably well while their bounding boxes are still unstable. To fix it, the authors build a dual-source framework: a cached, detector-independent reference pool of boxes and masks produced by UPN followed by SAM2, plus the current detector's own predictions. On top of this they propose FDSU, which fuses localization agreement between the two sources with classification confidence; OCDS, which does mask-guided diversity sampling in DINOv2 feature space while penalizing fragmented proposals; and DSBS, which swaps noisy detector boxes for the reference boxes during early annotation. The paper claims that on DIOR, HRSC2016, DOTAv2, and FAIR1M this combination matches or beats prior active learning methods under most annotation budgets, with the clearest gains in the low-budget cold-start regime. If right, it means fewer images need to be labeled and less manual box correction is needed to reach usable detection performance.

Core claim

The central claim is that a foundation-model-collaborative active learning loop, in which a category-agnostic UPN+SAM2 pipeline serves as a stable spatial reference while the detector supplies category and confidence signals, yields better sample selection and cheaper cold-start annotation than relying on the detector's own predictions alone. The paper asserts that FDSU+OCDS is superior or comparable to existing active learning methods under most annotation budgets across four remote sensing benchmarks, and that the DSBS module substantially improves early-stage Recall@75 at 20-30 labeled images (roughly 57%/61% mAP with DSBS vs 22%/45% without). The authors present this as evidence that loc

What carries the argument

The load-bearing mechanism is the SA/OD dual-source alignment. The SA-source is a one-time cached pool of category-agnostic boxes from UPN, refined through SAM2 masks into tight boxes and binary masks; the OD-source is the current-round detector's boxes and class confidences. Hungarian matching with an IoU-based cost (Eq. 1) pairs the two sets; FDSU fuses the matching agreement with classification confidence through a harmonic mean (Eq. 2) to score object-level uncertainty, aggregated by top-K per image. OCDS then builds mask-guided DINOv2 object embeddings, scores novelty against class prototypes, and subtracts a fragmentation penalty; DSBS replaces low-quality detector boxes by matched SA

Load-bearing premise

The load-bearing premise is that UPN+SAM2 boxes are a stable, detector-independent localization reference; if they are false positives or poorly aligned on dense, rotated small objects, both FDSU's matching and DSBS's box replacement inject the localization noise they are meant to remove.

What would settle it

Measure per-image annotator time or mouse-edit counts on DOTAv2 with DSBS enabled versus disabled; if no reduction in correction effort appears despite improved Recall@75, the annotation-cost claim collapses. A second check: jitter the cached SA boxes randomly and rerun FDSU+OCDS; if low-budget mAP barely changes, the SA-source is not the operative mechanism.

Watch this falsifier. Get emailed when new claim-graph text bears on it.

If this is right

  • Low-budget active learning on remote sensing object detection can be made much more sample-efficient; the strongest gains are reported at 10-50 labeled images.
  • A one-time offline UPN+SAM2 pass can be cached and reused in every active learning round, so the foundation-model inference cost is amortized.
  • DSBS provides a concrete semi-automatic annotation protocol: geometry from the reference source, labels from the detector, with unmatched boxes handled explicitly.
  • The framework transfers across diverse remote sensing benchmarks, including dense small objects (DOTAv2) and fine-grained categories (FAIR1M), suggesting the mechanism is not dataset-specific.
  • Ablations indicate that FDSU and OCDS contribute complementary gains; pairing either with a weak diversity strategy (CoreSet, DivProto) leaves much of the benefit unrealized.

Where Pith is reading between the lines

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

  • The paper never directly measures annotator time or correction effort; the claim that DSBS 'reduces manual burden' is inferred from Recall/mAP gains, so a human-in-the-loop timing study is the direct test.
  • The same dual-source recipe should transfer to other dense-object domains with a category-agnostic proposal model plus a mask generator, such as medical imaging or aerial video, where detector-only confidence is equally unreliable in early rounds.
  • One robustness gap the authors leave implicit: the SA-source is trusted equally everywhere, so a per-region or per-object trust weighting could prevent false UPN+SAM2 proposals from contaminating FDSU and DSBS on hard scenes.
  • DSBS scheduling via global matching quality (Eq. 10) could be refined to per-image or per-class switching thresholds, a direction the paper flags only as future work.
Share X Bluesky LinkedIn Reddit HN

Editorial analysis

A structured set of objections, weighed in public.

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

Referee Report

5 major / 6 minor

Summary. The paper proposes a foundation-model-assisted active learning and semi-automatic annotation framework for remote sensing object detection. It introduces a dual-source mechanism: an offline-cached 'SA-source' built by UPN proposals refined with SAM2 masks, and an online 'OD-source' from the current detector. Three components are added: FDSU (uncertainty from Hungarian matching between SA and OD boxes plus classification confidence), OCDS (DINOv2 mask-pooled object embeddings with fragmentation suppression), and DSBS (switching noisy detector boxes to SA boxes in cold-start rounds). Experiments on DIOR, HRSC2016, DOTAv2, and FAIR1M compare against Entropy, CoreSet, PPAL, DivProto, and AL4FM, with ablations on DIOR and a foundation-pipeline quality analysis in Table II.

Significance. The core idea—using a reusable, detector-independent localization source to stabilize cold-start active learning—is timely and practically motivated, and the one-time offline caching of SA-source proposals is a sensible design choice. If the reported gains are reproducible, the framework would be a useful step toward reducing annotation cost in remote sensing object detection. The paper also proposes a concrete mechanism (DSBS) to reduce box-editing burden, which is rarely evaluated in active learning papers. However, the current experimental evidence is not yet sufficient: the main comparison lacks error bars, the only tabulated ablation contains a non-monotonic baseline curve, the DSBS section mixes up Recall@100 and mAP, the SA-source is validated only by recall, and no direct measurement of manual annotation burden is provided. These gaps are load-bearing for the central claims of superior sample efficiency and reduced annotation cost.

major comments (5)
  1. [IV-B / Table I] Fig. 2, the central comparison across the four datasets, is shown without error bars, confidence bands, or the number of seeds. Active-learning pipelines are stochastic in initial labeled selection, detector training, and candidate sampling, so single-run differences of 1–3 mAP ('superior or comparable') are not interpretable. Table I, the only tabulated ablation, contains a red flag: the Entropy row gives mAP 0.5±0.3 at 10 labels and 0.2±0.1 at 50 labels, a non-monotonic drop that suggests a bug, unstable training, or very high variance. Please provide multi-seed statistics for all methods and investigate this baseline.
  2. [IV-B, DSBS analysis] The text says 'We evaluate using Recall@100 with a stricter IoU threshold of 75%' but then reports 'mAP' numbers (57%/61%, etc.). Recall@100 and mAP are different metrics; the text must specify which one is plotted in Fig. 3. In addition, the protocol is not described: are the 20/30 annotated samples selected by the AL loop, or is this a separate training run on a fixed small set? Without this information, the claim that DSBS 'substantially improves cold-start performance' cannot be verified.
  3. [III-A / III-D / Table II] The SA-source is characterized only by proposal recall and an undefined 'AvgΔIoU' column. No precision or object-level localization error relative to ground truth is reported, and there is no comparison of SA-box quality with detector-box quality on matched objects. This matters because FDSU's uncertainty signal (Eq. 2) and DSBS's replacement rule (Eq. 8) both assume SA boxes are a stable geometric reference. On DOTAv2, UPN+SAM2 recall at IoU 0.7 is only 58.8% (Top-200), leaving many objects without an SA box; how these unmatched objects are handled in DSBS is not analyzed. Please define AvgΔIoU, report SA precision/localization accuracy, and add a sensitivity analysis under SA-source degradation.
  4. [III-D / IV-B] The paper motivates DSBS by reducing the manual burden of box refinement, but no direct measurement of annotator effort appears anywhere. mAP or Recall gains are at best indirect proxies. Please add a user study or a quantitative proxy for editing cost—e.g., the fraction of displayed boxes that require correction, the average IoU of displayed boxes with the annotator-accepted box, or the number of mouse/keyboard edits per image—and report it under low-budget conditions.
  5. [III-C] In OCDS, the novelty of a candidate object is computed against prototypes of the class predicted by the detector (ĉ_j). In the cold-start regime the detector's classification is itself noisy; if ĉ_j is wrong, the object is measured against the wrong class prototypes and the coverage score is biased. Please analyze sensitivity to label noise or define a class-agnostic novelty fallback.
minor comments (6)
  1. [I] The module is first called 'Dual-Source Box Switching (DBS)' in the Introduction and later 'DSBS' throughout; please use one abbreviation consistently.
  2. [Table II] The 'AvgΔIoU' column is not defined. Add a formula or a description in the caption so the reader can interpret the values.
  3. [References] Reference [9] is cited as UPN, but the listed reference is 'ChatRex: Taming multimodal LLM...' The actual UPN paper appears to be missing; please correct the citation.
  4. [Figures 2-3] The figures have no visible axis labels, legends, or error bars in the provided text. Add these, and consider showing per-seed points for the main comparison.
  5. [Full text] There are minor typos, e.g., 'and and' in the Introduction, and the expectation notation E[·] in Eq. (10) is introduced without defining the distribution over images and matchings.
  6. [Method / Reproducibility] No code or detailed hyperparameter table is provided. Given the large number of hyperparameters (θ, Ku, Kod, K, Kd, p, λ, τsw, τa, τb, ρmin, ρmax, τupn, τcls, τloc), a supplementary configuration table would greatly aid reproducibility.

Circularity Check

0 steps flagged

No significant circularity; the central claims are benchmarked externally and the SA-source is validated against ground truth.

full rationale

The claimed contributions (FDSU, OCDS, DSBS) are constructive proposals, not results derived from inputs that already contain the asserted outcome. FDSU (Eq. 2) combines an externally computed SA-source (UPN+SAM2) with current detector outputs; neither term is defined in terms of the final selection outcome, and the localization agreement is measured against category-agnostic reference boxes rather than the evaluation labels. OCDS uses detector-predicted class labels to route candidate objects to prototypes built from the labeled set; this is a possible source of noise or bias, but it does not make the coverage score equal to the detector's predictions by construction. DSBS (Eq. 8) substitutes SA boxes for detector boxes, and its benefit is measured by training on the substituted boxes and testing on held-out ground truth, which is an external benchmark. Table II evaluates UPN/SAM2 recall against ground-truth boxes, so the foundation-pipeline quality claim has independent empirical support rather than being a self-evaluation by definition. The self-citations (e.g., [37], [38]) appear in related-work positioning and are not load-bearing for the method's derivation. Concerns about SA-source precision on dense/rotated objects and about self-training label noise are correctness/robustness questions, not circularity.

Axiom & Free-Parameter Ledger

10 free parameters · 6 axioms · 0 invented entities

The framework rests on several unverified domain assumptions about foundation-model boxes and features, and on a large set of hand-set hyperparameters whose values and sensitivity are not reported. No new physical or conceptual entities are introduced beyond framework components.

free parameters (10)
  • theta (candidate expansion factor θ)
    Selects nθ most uncertain images before diversity sampling (Sec. III-B.4); not reported beyond θ>1.
  • Top-Ku (UPN proposal retention)
    Controls SA-source recall/cost; the pre-filtering Top-Ku is described but its value is not given.
  • Top-Kod (detector box retention)
    Limits OD-source boxes after NMS; exact value not reported.
  • K (candidate objects per image limit) = 30-100 (example)
    Limits candidates in OCDS; K∈[30,100] stated as a typical range but exact per-dataset value is not given.
  • Kd (Top-Kd aggregation count)
    Number of top object novelties used in image-level coverage score; not specified.
  • p (prototypes per class) = 5
    k-means++ cluster count for class prototypes in Sec. III-C.2; fixed to p=5 in text.
  • lambda (fragmentation suppression weight)
    Weight in D(I)=Cov(I)-λ Frag(I); controls the selection trade-off and no value or sensitivity analysis is reported.
  • tau_sw (DSBS switching IoU threshold)
    Threshold in Eq. 8 deciding when an OD box is replaced by an SA box; no value reported.
  • tau_a / tau_b / rho_min / rho_max (geometric filters)
    Small-area and aspect-ratio filters in Sec. III-A and III-C; thresholds are hand-set and unreported.
  • tau_upn / tau_cls / tau_loc (DSBS scheduling thresholds)
    High-objectness, confidence pre-labeling, and adaptive DSBS scheduling thresholds; values not reported.
axioms (6)
  • domain assumption UPN+SAM2 provides a localization-stable, category-agnostic reference source independent of detector predictions.
    The entire FDSU and DSBS design depends on the SA-source being more reliable than cold-start detector boxes (Sec. III-A, III-B).
  • domain assumption IoU-based Hungarian matching identifies corresponding object instances across SA and OD sources.
    If the two sources produce geometrically different boxes for the same object, the localization agreement term in Eq. (2) becomes noisy (Sec. III-B.2).
  • domain assumption DINOv2 mask-guided embeddings organize remote sensing object appearances so that cosine novelty relative to class prototypes reflects annotation value.
    The coverage and novelty scoring in OCDS assumes semantic closeness in DINOv2 space is correlated with annotation usefulness (Sec. III-C).
  • domain assumption The current detector's predicted class labels are reliable enough to assign candidate objects to correct class prototypes during diversity sampling.
    OCDS uses OD-source predicted classes for prototype comparison, which can be wrong in the cold-start stage (Sec. III-C.2).
  • domain assumption mAP and Recall@100 improvements imply reduced manual annotation burden.
    The paper evaluates DSBS via detection metrics, not by measuring annotation time, correction counts, or annotator agreement (Sec. IV-B).
  • standard math Hungarian matching and k-means++ operate correctly in this setting.
    Used for one-to-one box alignment and prototype construction; no modification to the algorithms is proposed (Sec. III-B, III-C).

reviewed 2026-08-01 · how reviews work

0 comments
Cite this review

Pith. "Pith review of Foundation-Assisted Active Learning for Object Detection Annotation." pith.science (2026). https://pith.science/paper/N56HEYHD

@misc{pith2026260716671,
  author       = {Pith},
  title        = {Pith review of: Foundation-Assisted Active Learning for Object Detection Annotation},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/N56HEYHD}},
  note         = {Machine review of arXiv:2607.16671}
}
Share X Bluesky LinkedIn Reddit HN
read the original abstract

The annotation cost for remote sensing object detection is high, while existing active learning methods still face several challenges in object detection scenarios, including the coupling of localization and classification uncertainty, severe localization noise in the cold-start stage, and pseudo-diversity caused by high-recall candidate proposals. To address these issues, we propose a foundation-model-collaborative active learning and semi-automatic annotation framework for efficient construction of remote sensing object detection datasets. We build a dual-source mechanism consisting of a reference localization source (SA-source) based on UPN+SAM2 and a detector prediction source (OD-source), and further propose a Foundation-model-enhanced Dual-Source Uncertainty estimation to improve sample selection quality in the cold-start stage by jointly modeling localization consistency and classification confidence. Furthermore, we propose Object-Centric Diversity Sampling, which constructs object-level representations using DINOv2 features and SAM2 masks to improve sample coverage while suppressing pseudo-diversity. To address geometric noise in the semi-automatic annotation stage, we design Dual-Source Box Switching, which replaces noisy detector boxes with matched refined boxes from the SA-source, thereby reducing the manual burden of box refinement. Experiments on DIOR, HRSC2016, DOTAv2, and FAIR1M show that our method achieves superior or comparable results under most annotation budgets, with notably stronger cold-start sample efficiency in the low-budget regime.

Figures

Figures reproduced from arXiv: 2607.16671 by Arnold Zumbrun, Guoyu Lu, Jinchang Zhang, Jing Lin.

Figure 1
Figure 1. Figure 1: Foundation-model-collaborative active learning and semi-automatic annotation framework for remote sensing object detection. The left panel [PITH_FULL_IMAGE:figures/full_fig_p002_1.png] view at source ↗
Figure 2
Figure 2. Figure 2: Comparison of improvements under different annotation budgets [PITH_FULL_IMAGE:figures/full_fig_p006_2.png] view at source ↗
Figure 3
Figure 3. Figure 3: Cold-start results under different methods on HRSC2016 [24]. [PITH_FULL_IMAGE:figures/full_fig_p007_3.png] view at source ↗

discussion (0)

Sign in with ORCID, Apple, or X to comment. Anyone can read and Pith papers without signing in.

Reference graph

Works this paper leans on

38 extracted references · 8 linked inside Pith

  1. [1]

    Contextual diversity for active learning

    Sharat Agarwal, Himanshu Arora, Saket Anand, and Chetan Arora. Contextual diversity for active learning. In ECCV. Springer, 2020

    2020

  2. [2]

    Active learning via classifier impact and greedy selection for interactive image retrieval

    Leah Bar, Boaz Lerner, Nir Darshan, and Rami Ben-Ari. Active learning via classifier impact and greedy selection for interactive image retrieval. arXiv preprint arXiv:2412.02310 , 2024

    Pith/arXiv arXiv 2024

  3. [3]

    Active learning meets foundation models: fast remote sensing data annotation for object detection

    Marvin Burges, Philipe Ambrozio Dias, Carson Woody, Sarah Walters, and Dalton Lunga. Active learning meets foundation models: fast remote sensing data annotation for object detection. In ICCV, 2025

    2025

  4. [4]

    End-to-end object detec- tion with transformers

    Nicolas Carion, Francisco Massa, Gabriel Synnaeve, Nicolas Usunier, Alexander Kirillov, and Sergey Zagoruyko. End-to-end object detec- tion with transformers. In ECCV, pages 213–229. Springer, 2020

    2020

  5. [5]

    Emerging properties in self-supervised vision transformers

    Mathilde Caron, Hugo Touvron, Ishan Misra, Herv ´e J ´egou, Julien Mairal, Piotr Bojanowski, and Armand Joulin. Emerging properties in self-supervised vision transformers. In ICCV, 2021

    2021

  6. [6]

    Making your first choice: To address cold start problem in vision active learning

    Liangyu Chen, Yutong Bai, Siyu Huang, Yongyi Lu, Bihan Wen, Alan L Yuille, and Zongwei Zhou. Making your first choice: To address cold start problem in vision active learning. arXiv preprint arXiv:2210.02442, 2022

    Pith/arXiv arXiv 2022

  7. [7]

    Dav-gswt: Diffusion-active-view sampling for data-efficient gaussian splatting wang tiles

    Rong Fu, Jiekai Wu, Haiyun Wei, Yee Tan Jia, Yang Li, Xiaowen Ma, Wangyu Wu, and Simon Fong. Dav-gswt: Diffusion-active-view sampling for data-efficient gaussian splatting wang tiles. arXiv preprint arXiv:2602.15355, 2026

    Pith/arXiv arXiv 2026

  8. [8]

    T-rex2: Towards generic object detection via text-visual prompt synergy

    Qing Jiang, Feng Li, Zhaoyang Zeng, Tianhe Ren, Shilong Liu, and Lei Zhang. T-rex2: Towards generic object detection via text-visual prompt synergy. In ECCV, pages 38–57. Springer, 2024

    2024

  9. [9]

    Chatrex: Taming multimodal llm for joint perception and understanding

    Qing Jiang, Gen Luo, Yuqin Yang, Yuda Xiong, Yihao Chen, Zhaoyang Zeng, Tianhe Ren, and Lei Zhang. Chatrex: Taming multimodal llm for joint perception and understanding. arXiv preprint arXiv:2411.18363, 2024

    Pith/arXiv arXiv 2024

  10. [10]

    Object recognition consistency in regression for active detection

    Ming Jing, Zhilong Ou, Hongxing Wang, Jiaxin Li, and Ziyi Zhao. Object recognition consistency in regression for active detection. Machine Vision and Applications , 35(5):121, 2024

    2024

  11. [11]

    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 ICCV, 2023

    2023

  12. [12]

    Interactive multi-class tiny-object detection

    Chunggi Lee, Seonwook Park, Heon Song, Jeongun Ryu, Sanghoon Kim, Haejoon Kim, S ´ergio Pereira, and Donggeun Yoo. Interactive multi-class tiny-object detection. In CVPR, pages 14136–14145, 2022

    2022

  13. [13]

    Coreset selection for object detection

    Hojun Lee, Suyoung Kim, Junhoo Lee, Jaeyoung Yoo, and Nojun Kwak. Coreset selection for object detection. In CVPR, 2024

    2024

  14. [14]

    Object detection in optical remote sensing images: A survey and a new benchmark

    Ke Li, Gang Wan, Gong Cheng, Liqiu Meng, and Junwei Han. Object detection in optical remote sensing images: A survey and a new benchmark. ISPRS journal of photogrammetry and remote sensing , 159:296–307, 2020

    2020

  15. [15]

    Comae: Comprehensive attribute exploration for zero-shot hashing

    Yuqi Li, Qingqing Long, Yihang Zhou, Ran Zhang, Zhiyuan Ning, Zhihong Zhu, Yuanchun Zhou, Xuezhi Wang, and Meng Xiao. Comae: Comprehensive attribute exploration for zero-shot hashing. ICMR, 2025

    2025

  16. [16]

    3d plant root skeleton detection and extraction

    Jiakai Lin, Jinchang Zhang, Ge Jin, Wenzhan Song, Tianming Liu, and Guoyu Lu. 3d plant root skeleton detection and extraction. In 2025 IEEE/RSJ International Conference on Intelligent Robots and Systems (IROS), pages 3011–3017. IEEE, 2025

    2025

  17. [17]

    Keypoint detection and description for raw bayer images

    Jiakai Lin, Jinchang Zhang, and Guoyu Lu. Keypoint detection and description for raw bayer images. In 2025 IEEE International Conference on Robotics and Automation (ICRA) , pages 11736–11742. IEEE, 2025

    2025

  18. [18]

    Rt-detrv2: Improved baseline with bag-of-freebies for real- time detection transformer

    Wenyu Lv, Yian Zhao, Qinyao Chang, Kui Huang, Guanzhong Wang, and Yi Liu. Rt-detrv2: Improved baseline with bag-of-freebies for real- time detection transformer. arXiv preprint arXiv:2407.17140 , 2024

    Pith/arXiv arXiv 2024

  19. [19]

    Dinov2: Learning robust visual features without supervision

    Maxime Oquab, Timoth ´ee Darcet, Th ´eo Moutakanni, Huy V o, Marc Szafraniec, Vasil Khalidov, Pierre Fernandez, Daniel Haziza, Francisco Massa, Alaaeldin El-Nouby, et al. Dinov2: Learning robust visual features without supervision. arXiv preprint arXiv:2304.07193, 2023

    Pith/arXiv arXiv 2023

  20. [20]

    Sam 2: Segment anything in images and videos

    Nikhila Ravi, Valentin Gabeur, Yuan-Ting Hu, Ronghang Hu, Chai- tanya Ryali, Tengyu Ma, Haitham Khedr, Roman R ¨adle, Chloe Rol- land, Laura Gustafson, et al. Sam 2: Segment anything in images and videos. arXiv preprint arXiv:2408.00714 , 2024

    Pith/arXiv arXiv 2024

  21. [21]

    Active learning for convolutional neural networks: A core-set approach

    Ozan Sener and Silvio Savarese. Active learning for convolutional neural networks: A core-set approach. arXiv preprint, 2017

    2017

  22. [22]

    Oriane Sim ´eoni, Huy V V o, Maximilian Seitzer, Federico Baldas- sarre, Maxime Oquab, Cijo Jose, Vasil Khalidov, Marc Szafraniec, Seungeun Yi, Micha ¨el Ramamonjisoa, et al. Dinov3. arXiv preprint arXiv:2508.10104, 2025

    Pith/arXiv arXiv 2025

  23. [23]

    Fair1m: A benchmark dataset for fine-grained object recognition in high-resolution remote sensing imagery

    Xian Sun, Peijin Wang, Zhiyuan Yan, Feng Xu, Ruiping Wang, Wenhui Diao, Jin Chen, Jihao Li, Yingchao Feng, Tao Xu, et al. Fair1m: A benchmark dataset for fine-grained object recognition in high-resolution remote sensing imagery. ISPRS, 184, 2022

    2022

  24. [24]

    H-yolo: A single-shot ship detection approach based on region of interest preselected network

    Gang Tang, Shibo Liu, Iwao Fujino, Christophe Claramunt, Yide Wang, and Shaoyang Men. H-yolo: A single-shot ship detection approach based on region of interest preselected network. Remote Sensing, 12(24):4192, 2020

    2020

  25. [25]

    Curvature-adaptive consistency flow matching: Autonomous trajectory optimization via reinforcement learning, 2026

    Songtao Tian, Guhan Chen, Bohan Li, Jingyi Ma, and Zixiong Yu. Curvature-adaptive consistency flow matching: Autonomous trajectory optimization via reinforcement learning, 2026

    2026

  26. [26]

    Incorporating diversity and informativeness in multiple-instance active learning

    Ran Wang, Xi-Zhao Wang, Sam Kwong, and Chen Xu. Incorporating diversity and informativeness in multiple-instance active learning. IEEE transactions on fuzzy systems , 25(6):1460–1475, 2017

    2017

  27. [27]

    Alwod: Active learning for weakly-supervised object detection

    Yuting Wang, Velibor Ilic, Jiatong Li, Branislav Kisa ˇcanin, and Vladimir Pavlovic. Alwod: Active learning for weakly-supervised object detection. In ICCV, 2023

    2023

  28. [28]

    Entropy-based active learning for object detection with progressive diversity constraint

    Jiaxi Wu, Jiaxin Chen, and Di Huang. Entropy-based active learning for object detection with progressive diversity constraint. In CVPR, 2022

    2022

  29. [29]

    From points to coalitions: Hierarchical contrastive shapley values for prioritizing data samples

    Canran Xiao, Jiabao Dou, Zhiming Lin, Zong Ke, and Liwei Hou. From points to coalitions: Hierarchical contrastive shapley values for prioritizing data samples. In AAAI, 2026

    2026

  30. [30]

    Path matters: Unveiling geometric implicit bias via curvature-aware sparse view optimization

    Canran Xiao, Liaoyuan Fan, Yanbin Li, Jing Tang, and Peilai Yu. Path matters: Unveiling geometric implicit bias via curvature-aware sparse view optimization. In ICLR, 2026

    2026

  31. [31]

    Prototype-aligned federated soft-prompts for continual web personalization

    Canran Xiao and Liwei Hou. Prototype-aligned federated soft-prompts for continual web personalization. In WWW, 2026

    2026

  32. [32]

    Reversible primitive–composition alignment for continual vision–language learning

    Canran Xiao, Tianxiang Xu, Siyuan Ma, Yiyang Jiang, Haoyu Gao, and Yuhan Wu. Reversible primitive–composition alignment for continual vision–language learning. In ICLR, 2026

    2026

  33. [33]

    Plug and play active learning for object detection

    Chenhongyi Yang, Lichao Huang, and Elliot J Crowley. Plug and play active learning for object detection. In CVPR, 2024

    2024

  34. [34]

    Multi-class active learning by uncertainty sampling with diversity maximization

    Yi Yang, Zhigang Ma, Feiping Nie, Xiaojun Chang, and Alexander G Hauptmann. Multi-class active learning by uncertainty sampling with diversity maximization. IJCV, 113(2):113–127, 2015

    2015

  35. [35]

    Learning loss for active learning

    Donggeun Yoo and In So Kweon. Learning loss for active learning. In CVPR, 2019

    2019

  36. [36]

    Multiple instance active learning for object detection

    Tianning Yuan, Fang Wan, Mengying Fu, Jianzhuang Liu, Songcen Xu, Xiangyang Ji, and Qixiang Ye. Multiple instance active learning for object detection. In CVPR, 2021

    2021

  37. [37]

    Adaptive event stream slicing for open-vocabulary event-based object de- tection via vision-language knowledge distillation

    Jinchang Zhang, Zijun Li, Jiakai Lin, and Guoyu Lu. Adaptive event stream slicing for open-vocabulary event-based object de- tection via vision-language knowledge distillation. arXiv preprint arXiv:2510.00681, 2025

    arXiv 2025

  38. [38]

    Vision-language embodiment for monocular depth estimation

    Jinchang Zhang and Guoyu Lu. Vision-language embodiment for monocular depth estimation. In Proceedings of the Computer Vision and Pattern Recognition Conference , pages 29479–29489, 2025

    2025

This paper was first reviewed by deepseek-v4-flash on August 1, 2026.