Pith. sign in

REVIEW 5 major objections 5 minor 48 references

FlexiCrackNet: A Flexible Pipeline for Enhanced Crack Segmentation with General Features Transfered from SAM

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

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

Pith's one-line read FlexiCrackNet claims that a frozen EdgeSAM encoder, fused into a U-Net through gated attention, beats twelve crack-segmentation methods on DeepCrack and transfers zero-shot to CFD and Crack500 while using fewer parameters and less…

desk verdict Solid engineering result with a good zero-shot protocol, but the paper's story that frozen SAM features are the magic ingredient is not backed by its own ablation—review it, but push for the missing baselines. read the letter →

arxiv 2501.18855 v2 pith:JPPTVJCS submitted 2025-01-31 cs.CV

Xinlong Wan , Xiaoyan Jiang , Guangsheng Luo , Ferdous Sohel , Jenqneng Hwang This is my paper · ORCID
classification cs.CV
keywords cracksegmentationEdgeSAMSegmentAnythingModelzero-shotgeneralizationgatedattentionmechanismfeaturefusionencoder-decoderarchitecturepavementdetection
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

FlexiCrackNet proposes to reuse the frozen encoder of EdgeSAM, a lightweight CNN-based version of the Segment Anything Model, as a generic feature extractor for crack segmentation, then fuse those general features into a U-Net that is trained on crack images. The authors report that the resulting pipeline beats twelve comparison methods on the DeepCrack test set and in zero-shot transfer to the CFD and Crack500 datasets, while using fewer parameters and shorter inference time than the previous best crack-specific model. If the results are right, specialized segmentation models can inherit general visual knowledge from large pre-trained encoders without fine-tuning the foundation model or accepting its fixed 1024x1024 input resolution. That matters because foundation-model adaptation for small, specialized datasets is usually expensive and memory-hungry.

What carries the argument

The load-bearing mechanism is IGAM, the information-interaction gated attention mechanism. At each of five encoder stages, the crack-specific feature map from the U-Net branch and the corresponding generic feature map from the frozen EdgeSAM encoder are resized and channel-aligned by a scaling module, concatenated, and passed through an information interaction module that emits two attention masks: one that selects meaningful regions in the general features and one that enhances relevant details in the crack features. The masked maps are added together and combined with a residual connection that preserves the original crack-specific features. IGAM's job is to let the decoder use the generic prior without letting non-crack textures and semantics from the general encoder contaminate the segmentation.

What would settle it

Run the same DeepCrack training at 512x512 with a fine-tuned SAM-Adapter or fine-tuned EdgeSAM and evaluate zero-shot on CFD; if that baseline reaches or exceeds FlexiCrackNet's reported F1 of 54.48, the claim that frozen encoder features are the better transfer path would be contradicted.

Watch

Extended reading notes

Core claim

The paper's central claim is that task-agnostic features extracted by EdgeSAM's frozen CNN encoder, never fine-tuned on cracks, are a better prior for crack segmentation than training an encoder from scratch or adapting SAM at its fixed resolution. On the DeepCrack test set, FlexiCrackNet reports F1 82.88, IoU 71.33, and Dice 82.25, surpassing CrackMamba by 1.47, 2.56, and 2.00 points. On zero-shot evaluation over the CFD dataset it reports F1 54.48, IoU 39.15, and Dice 53.83, surpassing CrackMamba by 4.50, 4.75, and 4.58 points, and on zero-shot Crack500 validation it reports F1 56.85 versus CMTFNet's 48.76. The ablation attributes the gain to the IGAM fusion: replacing IGAM with plain concatenation drops F1 from 82.88 to 78.78 on DeepCrack and from 54.48 to 35.99 on CFD, while the bare U-Net baseline sits at 79.64 and 37.48. The pipeline also reports 25.63M parameters and 22.09 ms inference time, compared with CrackMamba's 75.95M parameters and 84.73 ms.

Load-bearing premise

The load-bearing premise is that the frozen EdgeSAM encoder's generic features, extracted without any crack-specific fine-tuning, are a better and more transferable source of visual prior for cracks than features from models trained or fine-tuned on crack data; the paper's evidence for this is its own reported experiments, and it does not include a direct fine-tuned-SAM comparison to test it.

Editorial extensions

If this is right

  • On the three benchmark settings the paper tests, DeepCrack supervised plus CFD and Crack500 zero-shot, FlexiCrackNet reports the highest F1, IoU, and Dice among the twelve compared methods.
  • Because only EdgeSAM's CNN encoder is used and its transformer decoder is dropped, the 1024x1024 input constraint of EdgeSAM no longer applies to the pipeline, which is what the paper calls flexible input resolution.
  • The model is smaller and faster than the leading crack-specific baseline, with 25.63M parameters and 22.09 ms inference versus 75.95M and 84.73 ms for CrackMamba, while reporting higher accuracy on both tested domains.
  • The IGAM ablation indicates that how generic features are fused matters: naive concatenation can hurt performance, while the gated attention recovers and exceeds the U-Net baseline.

Reading between the lines

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

  • Not tested in the paper: a direct comparison against an actual fine-tuned SAM baseline such as SAM-Adapter. If frozen features match or beat fine-tuning under the same schedule, the no-fine-tuning transfer route would be established; without that comparison, the advantage over the pre-training-plus-fine-tuning paradigm is inferred rather than demonstrated.
  • Not tested in the paper: input sizes other than 512x512. The claimed arbitrary-resolution flexibility follows from removing EdgeSAM's decoder, but all experiments resize inputs to 512x512, so a resolution sweep would be needed to confirm the design claim.
  • The fusion mechanism is domain-agnostic, so the same frozen-encoder-plus-gated-fusion recipe could plausibly transfer to other fine-grained segmentation tasks with small datasets, such as medical or industrial defect segmentation; the paper gestures at this broader applicability in its conclusion but does not test it.
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 FlexiCrackNet, a crack segmentation pipeline built on a U-Net encoder-decoder augmented with a frozen EdgeSAM encoder used as a generic feature extractor. The two feature streams are fused at five encoder stages by an information-interaction gated attention mechanism (IGAM) that generates attention masks for both general and crack-specific features, with a residual connection preserving the crack-specific stream. Training uses a combination of BCE and Dice losses. The method is evaluated on the DeepCrack training/test split and in a zero-shot protocol on the CFD and Crack500 datasets, comparing against twelve baselines and reporting F1, IoU, Dice, parameters, inference time, and GFLOPs. The authors report state-of-the-art results on DeepCrack and large zero-shot gains on CFD and Crack500, and claim flexibility with respect to input resolution and computational efficiency.

Significance. If the claims are confirmed, FlexiCrackNet would be a practically useful way to inject foundation-model priors into crack segmentation while avoiding the overhead of full SAM fine-tuning. The evaluation protocol is sound in its use of genuinely held-out datasets for zero-shot assessment, and the reported improvements are consistent across three datasets and multiple baselines. The authors also make the efficiency comparison transparent by reporting parameters, inference time, and GFLOPs. However, the central mechanistic claim that frozen EdgeSAM features are the source of the improvement is not isolated by the experiments, and the arbitrary-resolution claim is not tested. The lack of code release and single-run results additionally limit verification. The work is a reasonable engineering contribution, but the evidence currently supports a more modest claim than the one made in the abstract.

major comments (5)
  1. [§IV-F, Table III] The central claim that frozen EdgeSAM general features drive the improvement is not isolated. In Table III, replacing IGAM with plain concatenation ('Ours (Concat)') yields DeepCrack F1 78.78, below the U-Net baseline's 79.64, so the frozen EdgeSAM features are not helpful on their own; only the learned IGAM restores performance. Since Table I shows that Ours uses 25.63M parameters and 196.18 GFLOPs versus U-Net's 18.64M parameters and 101.99 GFLOPs, the +3.24 F1 gain over U-Net could plausibly come from added capacity and compute rather than from general feature transfer. A capacity-matched U-Net with comparable parameters and FLOPs should be evaluated to separate these effects.
  2. [§II-B, §IV-D] The paper argues that the proposed paradigm is preferable to the 'pre-training + fine-tuning' approach for SAM, but no SAM fine-tuning baseline is compared. SAM-Adapter is cited in Section II-B, and Zhou et al. [23] is described as adapting SAM for road crack segmentation, yet neither a SAM-Adapter nor a fine-tuned EdgeSAM variant appears in Tables I or II. Without such a comparison, the claimed advantage over the fine-tuning paradigm is unsubstantiated.
  3. [§I, §III-B, §IV-A] The claim of flexibility to 'diverse input resolutions' and 'arbitrary sizes' is not experimentally supported. Section IV-A states that all input images are resized to 512×512 pixels, and all experiments in Tables I and II use that resolution. No experiment varies the input resolution, so the decoupling of EdgeSAM from its fixed 1024×1024 input is demonstrated only architecturally, not empirically.
  4. [§IV-D, Tables I and II] All reported results are from single runs with no error bars or statistical significance tests. Some of the headline margins are small, for example 1.47 F1 over CrackMamba on DeepCrack and 2.45 F1 over CrackMamba on the Crack500 test set, so it is unclear whether these differences are robust to training variance. Reporting mean and standard deviation over at least three runs, or providing significance tests, would strengthen the central claim.
  5. [§IV-D, 'Efficiency comparison'] The statement that the method introduces only 'minimal additional computational cost' is misleading when compared to the U-Net baseline. Table I shows inference time increases from 4.59 ms for U-Net to 22.09 ms for FlexiCrackNet, and GFLOPs increase from 101.99 to 196.18. The efficiency advantage is real only relative to heavier transformer-based or SAM-based models such as CrackMamba; the comparison against the from-scratch U-Net should be stated more carefully.
minor comments (5)
  1. [Title] The word 'Transfered' in the title should be spelled 'Transferred'.
  2. [§IV-B] The dataset referred to as 'Cityforest' is commonly known as the Crack Forest Dataset (CFD); the name should be corrected for consistency with the literature.
  3. [Fig. 7] The figure caption lists EMCAD twice, as both (c) and (e); the second instance appears to be a typo.
  4. [References] There are two different DeepCrack references, [38] and [46], with different authors and years; the in-text citations should disambiguate which work is the dataset and which is the segmentation method.
  5. [§IV-A] The experimental configuration states that all models use 'a fixed loss function', but the paper does not specify whether this is the same BCE+Dice loss for all baselines or each baseline's original loss; clarifying this would strengthen the fairness of the comparison.

Circularity Check

0 steps flagged · score 0.0 of 10

No significant circularity: the zero-shot evaluations are genuinely held out and the frozen EdgeSAM features are externally pre-trained, so no fitted parameter or self-citation chain is equivalent to the claimed results.

full rationale

FlexiCrackNet's central claim is that frozen EdgeSAM encoder features, fused through a learned gating module, improve crack segmentation and zero-shot generalization. The derivation chain is experimentally grounded rather than circular. The model is trained only on the DeepCrack training split and evaluated on the DeepCrack test split, the full CFD dataset, and Crack500, none of which are used to fit any parameter of the proposed pipeline. The frozen EdgeSAM encoder is pre-trained externally (SAM/EdgeSAM) and is not fine-tuned on the target datasets, so the general-feature premise is an external input rather than a definitional restatement of the reported metrics. The ablation in Table III compares U-Net, EdgeSAM, concatenation, and IGAM; the fact that simple concatenation performs below the U-Net baseline does not indicate circularity, only that the learned fusion is necessary for the reported gains. The only apparent self-citation is reference [31] (TV-Net, which includes author X. Jiang) in the related-work discussion of encoder-decoder frameworks; it is peripheral and not load-bearing for the method's novelty or the reported numbers. Some claims, such as adaptability to arbitrary input resolutions, are unsupported because all experiments use 512x512 inputs, and the efficiency claim is weakened by the model's roughly doubled GFLOPs relative to the U-Net baseline, but those are support or correctness gaps, not circularity. No equation, fitted parameter, or self-citation chain is equivalent by construction to the claimed results, so the paper receives a low circularity score.

Assumptions & free parameters 5 free parameters · 3 assumptions · 2 invented entities

The ledger captures the hand-chosen hyperparameters and the domain assumptions behind the transfer learning recipe. No physical constants or external datasets are introduced beyond EdgeSAM's pre-trained weights, which are treated as an external input.

free parameters (5)
  • Initial learning rate = 0.0003
    Chosen by hand with cosine decay; not derived or ablated.
  • Batch size = 2
    Set by memory constraints; not varied.
  • Number of epochs = 100
    Stopped after a fixed schedule; no early stopping analysis.
  • Input resolution = 512x512
    All experiments use this resolution despite the claimed arbitrary-size flexibility.
  • Loss weights = BCE and Dice with equal weight
    No ablation on the loss combination is presented.
assumptions (3)
  • domain assumption Frozen CNN encoders pre-trained on large datasets provide transferable features for downstream tasks.
    Core premise of the method, supported by prior transfer learning literature but not by a dedicated experiment isolating this effect.
  • domain assumption The EdgeSAM encoder's multi-scale feature maps remain meaningful at the 512x512 input resolution used in all experiments.
    The paper claims arbitrary input size support but only tests one resolution.
  • standard math BCE plus Dice loss is an effective objective for crack segmentation.
    Common practice in the field; not derived or systematically compared.
invented entities (2)
  • Information-interaction gated attention mechanism (IGAM)
    purpose: Fuses general prior features from EdgeSAM with crack-specific U-Net features using two attention masks and a residual connection.
    Introduced and evaluated only inside this paper; no external validation or theoretical justification beyond the reported gains.
  • Information interaction module (IIM)
    purpose: Generates the attention masks inside IGAM from the concatenated feature map.
    Same as IGAM; its behavior is only shown through the overall results.

how reviews work

0 comments
Cite this review

Pith. "Pith review of FlexiCrackNet: A Flexible Pipeline for Enhanced Crack Segmentation with General Features Transfered from SAM." pith.science (2026). https://pith.science/paper/JPPTVJCS

@misc{pith2026250118855,
  author       = {Pith},
  title        = {Pith review of: FlexiCrackNet: A Flexible Pipeline for Enhanced Crack Segmentation with General Features Transfered from SAM},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/JPPTVJCS}},
  note         = {Machine review of arXiv:2501.18855}
}
read the original abstract

Automatic crack segmentation is a cornerstone technology for intelligent visual perception modules in road safety maintenance and structural integrity systems. Existing deep learning models and ``pre-training + fine-tuning'' paradigms often face challenges of limited adaptability in resource-constrained environments and inadequate scalability across diverse data domains. To overcome these limitations, we propose FlexiCrackNet, a novel pipeline that seamlessly integrates traditional deep learning paradigms with the strengths of large-scale pre-trained models. At its core, FlexiCrackNet employs an encoder-decoder architecture to extract task-specific features. The lightweight EdgeSAM's CNN-based encoder is exclusively used as a generic feature extractor, decoupled from the fixed input size requirements of EdgeSAM. To harmonize general and domain-specific features, we introduce the information-Interaction gated attention mechanism (IGAM), which adaptively fuses multi-level features to enhance segmentation performance while mitigating irrelevant noise. This design enables the efficient transfer of general knowledge to crack segmentation tasks while ensuring adaptability to diverse input resolutions and resource-constrained environments. Experiments show that FlexiCrackNet outperforms state-of-the-art methods, excels in zero-shot generalization, computational efficiency, and segmentation robustness under challenging scenarios such as blurry inputs, complex backgrounds, and visually ambiguous artifacts. These advancements underscore the potential of FlexiCrackNet for real-world applications in automated crack detection and comprehensive structural health monitoring systems.

Figures

Figures reproduced from arXiv: 2501.18855 by the authors.

Figure 1
Figure 1. Crack segmentation training paradigms. (a) Traditional [PITH_FULL_IMAGE:figures/full_fig_p002_1.png] view at source ↗
Figure 2
Figure 2. The figure depicts feature maps generated from input [PITH_FULL_IMAGE:figures/full_fig_p004_2.png] view at source ↗
Figure 2
Figure 2. The first nine feature maps of the five stages in the [PITH_FULL_IMAGE:figures/full_fig_p005_2.png] view at source ↗
Figures from the paper (6 more)
Figure 3
Figure 3. Figure 3: The overall pipeline of FlexiCrackNet. The pipeline adopts an encoder-decoder structure for crack segmentation, where [PITH_FULL_IMAGE:figures/full_fig_p006_3.png]
Figure 4
Figure 4. Figure 4: The structure of IIM. The total loss function is a combination of BCE and Dice Losses: Ltotal = LBce + LDice. (4) IV. EXPERIMENTS A. Experimental Configuration To ensure a fair and consistent comparison across all methods, we standardize the training configuration for …
Figure 5
Figure 5. Figure 5: Visualization of samples from the DeepCrack test set. [PITH_FULL_IMAGE:figures/full_fig_p009_5.png]
Figure 6
Figure 6. Figure 6: Visualization of samples from the CFD dataset. [PITH_FULL_IMAGE:figures/full_fig_p010_6.png]
Figure 7
Figure 7. Figure 7: Visualization of samples from the Crack500 dataset. [PITH_FULL_IMAGE:figures/full_fig_p010_7.png]
Figure 8
Figure 8. Figure 8: Comparison with the baseline. IGAM outperforming Concat by 4.10% in F1 score, 4.99% in IoU, and 4.30% in Dice coefficient on DeepCrack, and by 18.49% in F1 score, 13.99% in IoU, and 17.98% in Dice coefficient on CFD. These results validate the superiority of IGAM’s fus…

Discussion (0). Continue with ORCID to comment.

Reference graph

Works this paper leans on

48 extracted references · 44 canonical work pages

  1. [23]

    Teaching segment- anything-model domain-specific knowledge for road crack segmentation from on-board cameras,

    W. Zhou, H. Huang, H. Zhang, and C. Wang, “Teaching segment- anything-model domain-specific knowledge for road crack segmentation from on-board cameras,” IEEE Transactions on Intelligent Transporta- tion Systems, vol. 25, no. 12, pp. 20 588–20 601, 2024

    work page 2024

  2. [1]

    Automation in road distress detection, diagnosis and treatment,

    X. Yang, J. Zhang, W. Liu, J. Jing, H. Zheng, and W. Xu, “Automation in road distress detection, diagnosis and treatment,” Journal of Road Engineering, vol. 4, no. 1, pp. 1–26, 2024

    work page 2024

  3. [2]

    A universal multi-view guided network for salient object and camouflaged object detection,

    X. Song, P. Zhang, X. Lu, X. Hei, and R. Liu, “A universal multi-view guided network for salient object and camouflaged object detection,” IEEE Transactions on Circuits and Systems for Video Technology , vol. 34, no. 11, pp. 11 184–11 197, 2024

    work page 2024

  4. [3]

    Computer vision frame- work for crack detection of civil infrastructure—a review,

    D. Ai, G. Jiang, S.-K. Lam, P. He, and C. Li, “Computer vision frame- work for crack detection of civil infrastructure—a review,” Engineering Applications of Artificial Intelligence , vol. 117, p. 105478, 2023

    work page 2023

  5. [4]

    Erdunet: An efficient residual double- coding unet for medical image segmentation,

    H. Li, D.-H. Zhai, and Y . Xia, “Erdunet: An efficient residual double- coding unet for medical image segmentation,” IEEE Transactions on Circuits and Systems for Video Technology , vol. 34, no. 4, pp. 2083– 2096, 2024

    work page 2024

  6. [5]

    Pixel dif- ference convolutional network for rgb-d semantic segmentation,

    J. Yang, L. Bai, Y . Sun, C. Tian, M. Mao, and G. Wang, “Pixel dif- ference convolutional network for rgb-d semantic segmentation,” IEEE Transactions on Circuits and Systems for Video Technology , vol. 34, no. 3, pp. 1481–1492, 2024

    work page 2024

  7. [6]

    Boosting salient object detection with transformer-based asymmetric bilateral u-net,

    Y . Qiu, Y . Liu, L. Zhang, H. Lu, and J. Xu, “Boosting salient object detection with transformer-based asymmetric bilateral u-net,” IEEE Transactions on Circuits and Systems for Video Technology , vol. 34, no. 4, pp. 2332–2345, 2024

    work page 2024

  8. [7]

    Hybrid semantic segmentation for tunnel lining cracks based on swin transformer and convolutional neural network,

    Z. Zhou, J. Zhang, and C. Gong, “Hybrid semantic segmentation for tunnel lining cracks based on swin transformer and convolutional neural network,” Computer-Aided Civil and Infrastructure Engineering, vol. 38, no. 17, pp. 2491–2510, 2023

    work page 2023

Show all 48 references
  1. [8]

    Ctif-net: A cnn- transformer iterative fusion network for salient object detection,

    J. Yuan, A. Zhu, Q. Xu, K. Wattanachote, and Y . Gong, “Ctif-net: A cnn- transformer iterative fusion network for salient object detection,” IEEE Transactions on Circuits and Systems for Video Technology , vol. 34, no. 5, pp. 3795–3805, 2024

    2024

  2. [9]

    Small sample image segmen- tation by coupling convolutions and transformers,

    H. Qi, H. Zhou, J. Dong, and X. Dong, “Small sample image segmen- tation by coupling convolutions and transformers,” IEEE Transactions on Circuits and Systems for Video Technology, vol. 34, no. 7, pp. 5282– 5294, 2024

    2024

  3. [10]

    Ghostformer: Efficiently amal- gamated cnn-transformer architecture for object detection,

    X. Xie, D. Wu, M. Xie, and Z. Li, “Ghostformer: Efficiently amal- gamated cnn-transformer architecture for object detection,” Pattern Recognition, vol. 148, p. 110172, 2024

    2024

  4. [11]

    Long-short range adaptive transformer with dynamic sampling for 3d object detec- tion,

    C. Wang, J. Deng, J. He, T. Zhang, Z. Zhang, and Y . Zhang, “Long-short range adaptive transformer with dynamic sampling for 3d object detec- tion,” IEEE Transactions on Circuits and Systems for Video Technology, vol. 33, no. 12, pp. 7616–7629, 2023

    2023

  5. [12]

    A state-of-the- art survey of deep learning models for automated pavement crack segmentation,

    H. Gong, L. Liu, H. Liang, Y . Zhou, and L. Cong, “A state-of-the- art survey of deep learning models for automated pavement crack segmentation,” International Journal of Transportation Science and Technology, vol. 13, pp. 44–57, 2024

    2024

  6. [13]

    Automatic concrete defect detection and reconstruction by aligning aerial images onto semantic-rich building information model,

    J. Chen, W. Lu, and J. Lou, “Automatic concrete defect detection and reconstruction by aligning aerial images onto semantic-rich building information model,” Computer-Aided Civil and Infrastructure Engineer- ing, vol. 38, no. 8, pp. 1079–1098, 2023

    2023

  7. [14]

    Data augmentation in classification and segmentation: A survey and new strategies,

    K. Alomar, H. I. Aysel, and X. Cai, “Data augmentation in classification and segmentation: A survey and new strategies,” Journal of Imaging , vol. 9, no. 2, p. 46, 2023

    2023

  8. [15]

    Understanding and combating robust overfitting via input loss landscape analysis and regularization,

    L. Li and M. Spratling, “Understanding and combating robust overfitting via input loss landscape analysis and regularization,” Pattern Recogni- tion, vol. 136, p. 109229, 2023

    2023

  9. [16]

    Gpt-3: What’s it good for?

    R. Dale, “Gpt-3: What’s it good for?” Natural Language Engineering , vol. 27, no. 1, pp. 113–118, 2021

    2021

  10. [17]

    Learning transferable visual models from natural language supervision,

    A. Radford, J. W. Kim, C. Hallacy, A. Ramesh, G. Goh, S. Agarwal, G. Sastry, A. Askell, P. Mishkin, J. Clark et al., “Learning transferable visual models from natural language supervision,” in International conference on machine learning (ICML) , 2021, pp. 8748–8763

    2021

  11. [18]

    Segment anything,

    A. Kirillov, E. Mintun, N. Ravi, H. Mao, C. Rolland, L. Gustafson, T. Xiao, S. Whitehead, A. C. Berg, W.-Y . Loet al., “Segment anything,” in Proceedings of the IEEE/CVF International Conference on Computer Vision (ICCV), 2023, pp. 4015–4026

    2023

  12. [19]

    Detect any shadow: Segment anything for video shadow detection,

    Y . Wang, W. Zhou, Y . Mao, and H. Li, “Detect any shadow: Segment anything for video shadow detection,” IEEE Transactions on Circuits and Systems for Video Technology, vol. 34, no. 5, pp. 3782–3794, 2024

    2024

  13. [20]

    Sam-adapter: Adapting segment anything in underperformed scenes,

    T. Chen, L. Zhu, C. Deng, R. Cao, Y . Wang, S. Zhang, Z. Li, L. Sun, Y . Zang, and P. Mao, “Sam-adapter: Adapting segment anything in underperformed scenes,” in Proceedings of the IEEE/CVF International Conference on Computer Vision (ICCV) , 2023, pp. 3367–3375

    2023

  14. [21]

    Segment anything model for medical image analysis: an experimental study,

    M. A. Mazurowski, H. Dong, H. Gu, J. Yang, N. Konz, and Y . Zhang, “Segment anything model for medical image analysis: an experimental study,” Medical Image Analysis , vol. 89, p. 102918, 2023

    2023

  15. [22]

    Segment and caption anything,

    X. Huang, J. Wang, Y . Tang, Z. Zhang, H. Hu, J. Lu, L. Wang, and Z. Liu, “Segment and caption anything,” in Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR), 2024, pp. 13 405–13 417

    2024

  16. [24]

    Edgesam: Prompt-in-the- loop distillation for on-device deployment of sam,

    C. Zhou, X. Li, C. C. Loy, and B. Dai, “Edgesam: Prompt-in-the- loop distillation for on-device deployment of sam,” arXiv preprint arXiv:2312.06660, 2023

    2023 arXiv

  17. [25]

    Ecsnet: An accelerated real-time image segmentation cnn architecture for pavement crack detection,

    T. Zhang, D. Wang, and Y . Lu, “Ecsnet: An accelerated real-time image segmentation cnn architecture for pavement crack detection,” IEEE Transactions on Intelligent Transportation Systems , vol. 24, no. 12, pp. 15 105–15 112, 2023

    2023

  18. [26]

    Crackvit: a unified cnn-transformer model for pixel-level crack extraction,

    J. Quan, B. Ge, and M. Wang, “Crackvit: a unified cnn-transformer model for pixel-level crack extraction,” Neural Computing and Applica- tions, vol. 35, no. 15, pp. 10 957–10 973, 2023

    2023

  19. [27]

    A hybrid deep learning pavement crack semantic segmen- tation,

    Z. Al-Huda, B. Peng, R. N. A. Algburi, M. A. Al-antari, A.-J. Rabea, and D. Zhai, “A hybrid deep learning pavement crack semantic segmen- tation,” Engineering Applications of Artificial Intelligence , vol. 122, p. 106142, 2023

    2023

  20. [28]

    Fully convolutional networks for semantic segmentation,

    J. Long, E. Shelhamer, and T. Darrell, “Fully convolutional networks for semantic segmentation,” in Proceedings of the IEEE conference on computer vision and pattern recognition (CVPR) , 2015, pp. 3431–3440

    2015

  21. [29]

    U-net: Convolutional networks for biomedical image segmentation,

    O. Ronneberger, P. Fischer, and T. Brox, “U-net: Convolutional networks for biomedical image segmentation,” in in International Conference on Medical image computing and computer-assisted intervention (MIC- CAI), 2015, pp. 234–241

    2015

  22. [30]

    Segnet: A deep con- volutional encoder-decoder architecture for image segmentation,

    V . Badrinarayanan, A. Kendall, and R. Cipolla, “Segnet: A deep con- volutional encoder-decoder architecture for image segmentation,” IEEE transactions on pattern analysis and machine intelligence , vol. 39, no. 12, pp. 2481–2495, 2017

    2017

  23. [31]

    Tv-net: A structure- level feature fusion network based on tensor voting for road crack segmentation,

    W. Zheng, X. Jiang, Z. Fang, and Y . Gao, “Tv-net: A structure- level feature fusion network based on tensor voting for road crack segmentation,” IEEE Transactions on Intelligent Transportation Systems, vol. 25, no. 6, pp. 5743–5754, 2024

    2024

  24. [32]

    A nested unet with attention mechanism for road crack image segmentation,

    X. Fan, P. Cao, P. Shi, J. Wang, Y . Xin, and W. Huang, “A nested unet with attention mechanism for road crack image segmentation,” in 2021 IEEE 6th International Conference on Signal and Image Processing (ICSIP), 2021, pp. 189–193

    2021

  25. [33]

    Vision transformers for dense prediction,

    R. Ranftl, A. Bochkovskiy, and V . Koltun, “Vision transformers for dense prediction,” in Proceedings of the IEEE/CVF international conference on computer vision (ICCV) , 2021, pp. 12 179–12 188

    2021

  26. [34]

    Segformer: Simple and efficient design for semantic segmentation with transformers,

    E. Xie, W. Wang, Z. Yu, A. Anandkumar, J. M. Alvarez, and P. Luo, “Segformer: Simple and efficient design for semantic segmentation with transformers,” Advances in neural information processing systems , vol. 34, pp. 12 077–12 090, 2021

    2021

  27. [35]

    Vision transformer-based autonomous crack detection on asphalt and concrete surfaces,

    E. A. Shamsabadi, C. Xu, A. S. Rao, T. Nguyen, T. Ngo, and D. Dias-da Costa, “Vision transformer-based autonomous crack detection on asphalt and concrete surfaces,” Automation in Construction, vol. 140, p. 104316, 2022

    2022

  28. [36]

    A convolutional-transformer network for crack segmentation with boundary awareness,

    H. Tao, B. Liu, J. Cui, and H. Zhang, “A convolutional-transformer network for crack segmentation with boundary awareness,” in2023 IEEE International Conference on Image Processing (ICIP) , 2023, pp. 86–90

    2023

  29. [37]

    Parameter-efficient transfer 13 learning for nlp,

    N. Houlsby, A. Giurgiu, S. Jastrzebski, B. Morrone, Q. De Laroussilhe, A. Gesmundo, M. Attariyan, and S. Gelly, “Parameter-efficient transfer 13 learning for nlp,” in International conference on machine learning (ICML), 2019, pp. 2790–2799

    2019

  30. [38]

    Deepcrack: A deep hierarchical feature learning architecture for crack segmentation,

    Y . Liu, J. Yao, X. Lu, R. Xie, and L. Li, “Deepcrack: A deep hierarchical feature learning architecture for crack segmentation,” Neurocomputing, vol. 338, pp. 139–153, 2019

    2019

  31. [39]

    Automatic road crack detection using random structured forests,

    Y . Shi, L. Cui, Z. Qi, F. Meng, and Z. Chen, “Automatic road crack detection using random structured forests,” IEEE Transactions on Intel- ligent Transportation Systems, vol. 17, no. 12, pp. 3434–3445, 2016

    2016

  32. [40]

    Feature pyramid and hierarchical boosting network for pavement crack detection,

    F. Yang, L. Zhang, S. Yu, D. Prokhorov, X. Mei, and H. Ling, “Feature pyramid and hierarchical boosting network for pavement crack detection,” IEEE Transactions on Intelligent Transportation Systems , vol. 21, no. 4, pp. 1525–1535, 2020

    2020

  33. [41]

    Pyramid scene parsing network,

    H. Zhao, J. Shi, X. Qi, X. Wang, and J. Jia, “Pyramid scene parsing network,” in Proceedings of the IEEE conference on computer vision and pattern recognition (CVPR) , 2017, pp. 2881–2890

    2017

  34. [42]

    Encoder- decoder with atrous separable convolution for semantic image segmen- tation,

    L.-C. Chen, Y . Zhu, G. Papandreou, F. Schroff, and H. Adam, “Encoder- decoder with atrous separable convolution for semantic image segmen- tation,” in Proceedings of the European conference on computer vision (ECCV), 2018, pp. 801–818

    2018

  35. [43]

    Emcad: Efficient multi- scale convolutional attention decoding for medical image segmentation,

    M. M. Rahman, M. Munir, and R. Marculescu, “Emcad: Efficient multi- scale convolutional attention decoding for medical image segmentation,” in Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR), 2024, pp. 11 769–11 779

    2024

  36. [44]

    Multi-scale high-resolution vision transformer for semantic segmentation,

    J. Gu, H. Kwon, D. Wang, W. Ye, M. Li, Y .-H. Chen, L. Lai, V . Chandra, and D. Z. Pan, “Multi-scale high-resolution vision transformer for semantic segmentation,” in Proceedings of the IEEE/CVF conference on computer vision and pattern recognition (CVPR) , 2022, pp. 12 094– 12 103

    2022

  37. [45]

    Cmtfnet: Cnn and multiscale transformer fusion network for remote sensing image semantic segmentation,

    H. Wu, P. Huang, M. Zhang, W. Tang, and X. Yu, “Cmtfnet: Cnn and multiscale transformer fusion network for remote sensing image semantic segmentation,” IEEE Transactions on Geoscience and Remote Sensing, vol. 61, pp. 1–12, 2023

    2023

  38. [46]

    Deepcrack: Learning hierarchical convolutional features for crack detection,

    Q. Zou, Z. Zhang, Q. Li, X. Qi, Q. Wang, and S. Wang, “Deepcrack: Learning hierarchical convolutional features for crack detection,” IEEE transactions on image processing , vol. 28, no. 3, pp. 1498–1512, 2018

    2018

  39. [47]

    Deepcrackat: An effective crack segmentation framework based on learning multi-scale crack features,

    Q. Lin, W. Li, X. Zheng, H. Fan, and Z. Li, “Deepcrackat: An effective crack segmentation framework based on learning multi-scale crack features,” Engineering Applications of Artificial Intelligence , vol. 126, p. 106876, 2023

    2023

  40. [48]

    Topology-aware mamba for crack segmentation in structures,

    X. Zuo, Y . Sheng, J. Shen, and Y . Shan, “Topology-aware mamba for crack segmentation in structures,” Automation in Construction, vol. 168, p. 105845, 2024

    2024

Pith tools

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