Pith. sign in

REVIEW 3 major objections 5 minor 60 references

A Lightweight Group Multiscale Bidirectional Interactive Network for Real-Time Steel Surface Defect Detection

T0 review · 3 major / 5 minor · reviewed 2026-08-05 · deepseek-v4-flash

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

Pith's one-line read A 0.19-million-parameter detector built from group-wise multiscale modules with parameter-free cross-scale interaction claims competitive steel-defect accuracy at 1048 FPS on GPU and 16.53 FPS on CPU at 512 resolution.

desk verdict A plausible lightweight-detector architecture with a falsifiable efficiency claim; needs the measurement protocol and accuracy numbers before the speed headline can be believed. read the letter →

arxiv 2508.16397 v1 pith:VONZ4P2K submitted 2025-08-22 cs.CV cs.AI

Yong Zhang , Cunjian Chen , Qiang Gao , Yi Wang , Bin Fang This is my paper
classification cs.CVcs.AI
keywords steelsurfacedefectdetectionlightweightneuralnetworkreal-timeobjectmultiscalefeatureinteractioncross-scalefusiondepthwiseseparableconvolutionindustrialvisionGMBINet
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 is trying to establish that a very small neural network can perform real-time steel surface defect detection without sacrificing accuracy to larger models. Its proposed GMBINet uses group-wise multiscale modules whose computational cost does not grow with the number of scales, plus bidirectional cross-scale interaction that adds no extra parameters. The evidence is competitive accuracy on two defect-detection benchmarks and very high frame rates at 512 resolution with only 0.19 million parameters. A reader should care because factory-floor detection needs models that run on cheap CPUs and GPUs, and this paper claims a route to that at a fraction of typical model size.

What carries the argument

The central object is the Group Multiscale Bidirectional Interactive (GMBI) module. It extracts multiscale features in channel groups, so the number of scales does not multiply cost, and it combines a Bidirectional Progressive Feature Interactor (BPFI), which passes features up and down between scale levels, with an Element-Wise Multiplication-Summation (EWMS) operation that fuses those two directions without learned weights. Together these mechanisms are meant to provide cross-scale interaction at zero extra parameter cost, keeping the entire detector at 0.19 million parameters.

What would settle it

Reproduce GMBINet and the fastest baseline detectors at 512 resolution on the same GPU and CPU, with batch size 1 and identical inference engine and precision. If GMBINet cannot sustain close to 1048 FPS on GPU and 16.53 FPS on CPU while keeping the reported accuracy, or if the baselines reach similar speeds once identically optimized, the central real-time claim is refuted.

Watch

Extended reading notes

Core claim

The paper proposes GMBINet, a lightweight detector built from Group Multiscale Bidirectional Interactive (GMBI) modules. Each GMBI module groups channels and applies multiscale feature extraction within those groups so that adding scales does not multiply computation or parameters. It then uses a Bidirectional Progressive Feature Interactor (BPFI) and a parameter-free Element-Wise Multiplication-Summation (EWMS) operation to combine cross-scale information. On the SD-Saliency-900 and NRSD-MN datasets, the authors report competitive defect-detection accuracy while running at 1048 FPS on GPU and 16.53 FPS on CPU at 512 resolution, with only 0.19 million parameters, and they show that the learn

Load-bearing premise

The load-bearing premise is that the reported 1048 FPS and 16.53 FPS were measured under the same inference protocol—same GPU and CPU models, batch size, precision, and software optimizations—as the comparison methods, so the real-time advantage is fair and not a measurement artifact.

Editorial extensions

If this is right

  • A detector with 0.19 million parameters can match or approach the accuracy of much larger models on public steel-defect benchmarks, so real-time inline inspection no longer requires a datacenter GPU.
  • The parameter-free cross-scale interactor implies that richer feature fusion does not have to cost additional parameters, an option available to any lightweight detector.
  • At 1048 FPS, a single GPU can process live camera streams at full frame rate with headroom for preprocessing, making continuous production-line screening feasible.
  • At 16.53 FPS on CPU at 512 resolution, the model is usable on industrial PCs and edge boxes without dedicated accelerators.
  • Transfer to the NEU-CLS classification benchmark suggests the learned multiscale representation generalizes beyond bounding-box detection to other defect-recognition tasks.

Reading between the lines

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

  • The reported 1048 FPS is likely a best-case latency measurement; end-to-end deployment would add image capture, preprocessing, postprocessing, and communication, so practical throughput will be lower, though still plausibly real-time on GPU.
  • Because the EWMS interactor adds no parameters, it could be inserted into other lightweight backbones as a plug-in cross-scale fusion layer; testing this on non-steel tasks would reveal whether the mechanism itself, rather than the specific architecture, carries the gains.
  • The group-wise scale-agnostic design suggests a general recipe: split channels into groups instead of adding parallel branches when building multiscale networks; this could reduce cost in segmentation, remote sensing, or medical imaging where multiscale context matters.
  • A production test on unseen steel grades, different lighting, and higher-resolution images would show whether the benchmark accuracy advantage survives distribution shift, since defect datasets often under-represent real factory surface variety.
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

3 major / 5 minor

Summary. The paper proposes GMBINet, a lightweight architecture for real-time steel surface defect detection. The contribution is a Group Multiscale Bidirectional Interactive (GMBI) module that combines group-wise multiscale extraction, a Bidirectional Progressive Feature Interactor (BPFI), and a parameter-free Element-Wise Multiplication-Summation (EWMS) operation. The central empirical claim, stated in the abstract, is that GMBINet achieves competitive accuracy on SD-Saliency-900 and NRSD-MN, transfers to NEU-CLS classification, and delivers 1048 FPS on GPU and 16.53 FPS on CPU at 512 resolution with only 0.19 million parameters.

Significance. If the reported numbers hold under a fair and reproducible measurement protocol, GMBINet would be a significant result: a sub-0.2M-parameter detector running at 1048 FPS with competitive accuracy is far beyond typical lightweight detectors and would be practically valuable for industrial deployment. Strengths include the public dataset and code links, which enable independent verification, and the architectural design that avoids additional parameters via parameter-free operations. However, the significance is conditional because the abstract does not quantify the accuracy claim or provide the measurement conditions for the speed numbers, which are the main selling points.

major comments (3)
  1. [Abstract] The headline speed figures—1048 FPS on GPU and 16.53 FPS on CPU at 512 resolution—are stated without the measurement protocol needed to interpret them. Please specify the GPU and CPU models, batch size, precision (FP32/FP16/INT8), inference framework (PyTorch/TensorRT/ONNX), whether the reported number is latency or throughput, and whether post-processing such as NMS is included. Without these details, the numbers cannot be compared to baseline methods, and the central claim of a real-time advantage is not verifiable.
  2. [Abstract] The phrase 'competitive accuracy' is unquantified. No numerical accuracy metrics (e.g., mAP, F1, or precision/recall) are given for GMBINet or for the comparison methods on SD-Saliency-900 and NRSD-MN. Because the contribution is an accuracy–speed tradeoff, the reader needs the actual numbers and the conditions under which they were measured to judge whether the tradeoff is genuinely competitive. Please include quantitative results, or clearly refer to tables in the experimental section with full protocol details.
  3. [Abstract] The assertion of 'scale-agnostic computational complexity' is not defined or supported. The computational complexity of a convolutional network generally depends on input resolution through the spatial dimensions of feature maps. If the group-wise strategy makes FLOPs independent of resolution, a derivation or FLOPs formula is required. If the intended meaning is a weaker claim (e.g., 'nearly constant' or 'independent of the number of groups'), state it precisely. Otherwise the claim is misleading.
minor comments (5)
  1. [Abstract] The parameter count '0.19 M' should clarify whether it includes only trainable weights or also BatchNorm buffers; the convention matters for reproducibility.
  2. [Abstract] '512 resolution' is ambiguous; specify 512×512 input to avoid confusion with aspect ratio or minimum side.
  3. [Abstract] The datasets SD-Saliency-900, NRSD-MN, and NEU-CLS are named without definitions. Provide brief descriptions or citations at first mention.
  4. [Abstract] EWMS is introduced but the operations behind the acronym are not stated. A short parenthetical explanation would help the reader understand the parameter-free nature.
  5. [Abstract] The NEU-CLS evaluation is a classification task, whereas the main method is a detection network. Clarify how the model is adapted for classification (e.g., global pooling plus a linear head) to avoid confusion about transferability.

Circularity Check

0 steps flagged · score 0.0 of 10

No circularity identified; the paper's claims are empirical and benchmarked against external datasets.

full rationale

The paper is an architecture paper: it proposes GMBINet with group multiscale modules, bidirectional progressive feature interaction, and element-wise multiplication-summation, and evaluates the resulting network on external datasets (SD-Saliency-900, NRSD-MN, NEU-CLS) against external benchmarks. There is no derivation chain in which a predicted quantity is defined in terms of the quantity it is supposed to predict, no fitted parameter is renamed as a prediction, and no load-bearing result is justified solely by self-citation. The claimed FPS numbers and 0.19 M parameter count are empirical results that depend on measurement protocol, but concern about unstated GPU/CPU models, batch size, or precision is a reproducibility and validity issue, not circularity. The 'scale-agnostic computational complexity' assertion is a design property, not a result derived from the target metric. Since no specific equation or definition reduces to its own input, the appropriate finding is no significant circularity.

Assumptions & free parameters 1 free parameters · 3 assumptions · 3 invented entities

This is an empirical architecture paper. The central claim rests on trained weights (0.19M parameters), hand-chosen architecture hyperparameters not disclosed in the abstract, three public datasets whose splits and protocols are unspecified, and a speed-measurement protocol that is not described. There is no derivation in the abstract, so there are no fitted constants being relabeled as predictions. The main ledger concerns are the undisclosed hyperparameters and the unverifiable evaluation protocol.

free parameters (1)
  • Architecture hyperparameters (group count, per-group channels, GMBI stage depth, interaction order) = undisclosed in abstract
    The 0.19M parameter budget and the claimed scale-agnostic complexity depend on hand-chosen design choices that the abstract does not report; these choices directly determine the headline efficiency and cannot be audited from the abstract.
assumptions (3)
  • domain assumption Multibranch depthwise-separable-convolution designs are an appropriate base for lightweight defect detection
    The paper motivates GMBINet as an improvement over 'recent lightweight approaches' built on DSConv; this framing inherits the design space from prior literature rather than justifying it.
  • domain assumption The datasets SD-Saliency-900, NRSD-MN, and NEU-CLS measure real steel defect detection and classification capability
    All accuracy claims rest on these three benchmarks; the abstract does not address how well they represent production steel surfaces, and the train/test splits are not specified.
  • domain assumption The reported FPS values are representative of real deployment performance
    The 1048 FPS and 16.53 FPS figures presume a measurement protocol (hardware, batch size, precision, averaging) that the abstract does not describe; latency claims of this type are highly protocol-dependent.
invented entities (3)
  • GMBI module (Group Multiscale Bidirectional Interactive)
    purpose: Group-wise multiscale feature extraction claimed to keep computational complexity independent of the number of scales
    The module's value is evidenced only by the paper's own experiments; no external reproduction or independent analysis is cited, which is normal for an architecture paper but means the entity's claimed property is asserted, not independently verified.
  • BPFI (Bidirectional Progressive Feature Interactor)
    purpose: Cross-scale feature interaction by passing information progressively between detail levels in both directions
    Same support structure as GMBI: only in-paper experiments support the claim that this interactor improves accuracy without adding overhead.
  • EWMS (Element-Wise Multiplication-Summation)
    purpose: Parameter-free fusion operation that combines element-wise multiplication and summation to exchange cross-scale information with no added computation
    The claim that EWMS adds zero computational overhead is an architectural assertion that would require a FLOPs and latency accounting to verify; no such accounting is visible in the abstract.

how reviews work

0 comments
Cite this review

Pith. "Pith review of A Lightweight Group Multiscale Bidirectional Interactive Network for Real-Time Steel Surface Defect Detection." pith.science (2026). https://pith.science/paper/VONZ4P2K

@misc{pith2026250816397,
  author       = {Pith},
  title        = {Pith review of: A Lightweight Group Multiscale Bidirectional Interactive Network for Real-Time Steel Surface Defect Detection},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/VONZ4P2K}},
  note         = {Machine review of arXiv:2508.16397}
}
read the original abstract

Real-time surface defect detection is critical for maintaining product quality and production efficiency in the steel manufacturing industry. Despite promising accuracy, existing deep learning methods often suffer from high computational complexity and slow inference speeds, which limit their deployment in resource-constrained industrial environments. Recent lightweight approaches adopt multibranch architectures based on depthwise separable convolution (DSConv) to capture multiscale contextual information. However, these methods often suffer from increased computational overhead and lack effective cross-scale feature interaction, limiting their ability to fully leverage multiscale representations. To address these challenges, we propose GMBINet, a lightweight framework that enhances multiscale feature extraction and interaction through novel Group Multiscale Bidirectional Interactive (GMBI) modules. The GMBI adopts a group-wise strategy for multiscale feature extraction, ensuring scale-agnostic computational complexity. It further integrates a Bidirectional Progressive Feature Interactor (BPFI) and a parameter-free Element-Wise Multiplication-Summation (EWMS) operation to enhance cross-scale interaction without introducing additional computational overhead. Experiments on SD-Saliency-900 and NRSD-MN datasets demonstrate that GMBINet delivers competitive accuracy with real-time speeds of 1048 FPS on GPU and 16.53 FPS on CPU at 512 resolution, using only 0.19 M parameters. Additional evaluations on the NEU-CLS defect classification dataset further confirm the strong generalization ability of our method, demonstrating its potential for broader industrial vision applications beyond surface defect detection. The dataset and code are publicly available at: https://github.com/zhangyongcode/GMBINet.

Discussion (0). Sign in to comment.

Reference graph

Works this paper leans on

60 extracted references · 58 canonical work pages

  1. [1]

    write newline

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

  2. [2]

    MINet : Multiscale interactive network for real-time salient object detection of strip steel surface defects

    Kunye Shen, Xiaofei Zhou, and Zhi Liu. MINet : Multiscale interactive network for real-time salient object detection of strip steel surface defects. IEEE Trans. Ind. Inform. , 20(5):7842--7852, 2024

    work page 2024

  3. [3]

    EDRNet : Encoder--decoder residual network for salient object detection of strip steel surface defects

    Guorong Song, Kechen Song, and Yunhui Yan. EDRNet : Encoder--decoder residual network for salient object detection of strip steel surface defects. IEEE Trans. Instrum. Meas. , 69(12):9709--9719, 2020

    work page 2020

  4. [4]

    Saliency optimization from robust background detection

    Wangjiang Zhu, Shuang Liang, Yichen Wei, and Jian Sun. Saliency optimization from robust background detection. In Proc. IEEE Conf. Comput. Vis. Pattern Recognit. , pages 2814--2821, 2014

    work page 2014

  5. [5]

    A cluster-based wavelet feature extraction method and its application

    Gang Yu and Sagar V Kamarthi. A cluster-based wavelet feature extraction method and its application. Eng. Appl. Artif. Intell. , 23(2):196--202, 2010

    work page 2010

  6. [6]

    A lightweight and robust detection network for diverse glass surface defects via scale-and shape-aware feature extraction

    Huan Yu, Jin Wang, Jingru Yang, Yiming Liang, Zhihui Li, Zhan Wang, Haiyan He, Xusheng Zhang, and Guodong Lu. A lightweight and robust detection network for diverse glass surface defects via scale-and shape-aware feature extraction. Eng. Appl. Artif. Intell. , 153:110640, 2025

    work page 2025

  7. [7]

    A systematic review of deep learning approaches for surface defect detection in industrial applications

    Rasoul Ameri, Chung-Chian Hsu, and Shahab S Band. A systematic review of deep learning approaches for surface defect detection in industrial applications. Eng. Appl. Artif. Intell. , 130:107717, 2024

    work page 2024

  8. [8]

    A review on modern defect detection models using dcnns--deep convolutional neural networks

    Andrei-Alexandru Tulbure, Adrian-Alexandru Tulbure, and Eva-Henrietta Dulf. A review on modern defect detection models using dcnns--deep convolutional neural networks. J. Adv. Res. , 35:33--48, 2022

    work page 2022

Show all 60 references
  1. [9]

    Review of surface defect detection of steel products based on machine vision

    Bo Tang, Li Chen, Wei Sun, and Zhong-kang Lin. Review of surface defect detection of steel products based on machine vision. IET Image Proc. , 17(2):303--322, 2023

    2023

  2. [10]

    Dense attention-guided cascaded network for salient object detection of strip steel surface defects

    Xiaofei Zhou, Hao Fang, Zhi Liu, Bolun Zheng, Yaoqi Sun, Jiyong Zhang, and Chenggang Yan. Dense attention-guided cascaded network for salient object detection of strip steel surface defects. IEEE Trans. Instrum. Meas. , 71:1--14, 2021

    2021

  3. [11]

    Edge-aware multi-level interactive network for salient object detection of strip steel surface defects

    Xiaofei Zhou, Hao Fang, Xiaobo Fei, Ran Shi, and Jiyong Zhang. Edge-aware multi-level interactive network for salient object detection of strip steel surface defects. IEEE Access , 9:149465--149476, 2021

    2021

  4. [12]

    Very deep convolutional networks for large-scale image recognition

    Karen Simonyan and Andrew Zisserman. Very deep convolutional networks for large-scale image recognition. In Proc. Int. Conf. Learn. Represent. , pages 1--14, 2015

    2015

  5. [13]

    Deep residual learning for image recognition

    Kaiming He, Xiangyu Zhang, Shaoqing Ren, and Jian Sun. Deep residual learning for image recognition. In IEEE Conf. Comput. Vis. Pattern Recognit. , pages 770--778, 2016

    2016

  6. [14]

    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 Proc. IEEE Int. Conf. Comput. Vis. , pages 10012--10022, 2021

    2021

  7. [15]

    Mobilenets: Efficient convolutional neural networks for mobile vision applications

    Andrew G Howard. Mobilenets: Efficient convolutional neural networks for mobile vision applications. arXiv preprint arXiv:1704.04861 , 2017

    2017 arXiv

  8. [16]

    Shufflenet v2: Practical guidelines for efficient cnn architecture design

    Ningning Ma, Xiangyu Zhang, Hai-Tao Zheng, and Jian Sun. Shufflenet v2: Practical guidelines for efficient cnn architecture design. In Proc. Eur. Conf. Comput. Vis. , pages 116--131, 2018

    2018

  9. [17]

    Ghostnet: More features from cheap operations

    Kai Han, Yunhe Wang, Qi Tian, Jianyuan Guo, Chunjing Xu, and Chang Xu. Ghostnet: More features from cheap operations. In Proc. IEEE Conf. Comput. Vis. Pattern Recognit. , pages 1580--1589, 2020

    2020

  10. [18]

    SAMNet : Stereoscopically attentive multi-scale network for lightweight salient object detection

    Yun Liu, Xin-Yu Zhang, Jia-Wang Bian, Le Zhang, and Ming-Ming Cheng. SAMNet : Stereoscopically attentive multi-scale network for lightweight salient object detection. IEEE Trans. Image Process. , 30:3804--3814, 2021

    2021

  11. [19]

    An improved you only look once model for the multi-scale steel surface defect detection with multi-level alignment and cross-layer redistribution features

    Jianhang Huang, Xinliang Zhang, Lijie Jia, and Yitian Zhou. An improved you only look once model for the multi-scale steel surface defect detection with multi-level alignment and cross-layer redistribution features. Eng. Appl. Artif. Intell. , 145:110214, 2025

    2025

  12. [20]

    SRPCNet : Self-reinforcing perception coordination network for seamless steel pipes internal surface defect detection

    Hongshu Chen, Kechen Song, Wenqi Cui, Tianle Zhang, Yunhui Yan, and Jun Li. SRPCNet : Self-reinforcing perception coordination network for seamless steel pipes internal surface defect detection. IEEE Trans. Ind. Inform. , 21(1):950--958, 2025

    2025

  13. [21]

    CACS-YOLO : A lightweight model for insulator defect detection based on improved yolov8m

    Zhong Cao, Kaihong Chen, Junzuo Chen, Zhaohui Chen, and Man Zhang. CACS-YOLO : A lightweight model for insulator defect detection based on improved yolov8m. IEEE Trans. Instrum. Meas. , 73:1--10, 2024

    2024

  14. [22]

    SIA-net : Structural information awareness network based on normal samples for surface defect detection

    Qiurui Ma, Erhu Zhang, Yajun Chen, Jinghong Duan, and Linhao Shao. SIA-net : Structural information awareness network based on normal samples for surface defect detection. Eng. Appl. Artif. Intell. , 126:107131, 2023

    2023

  15. [23]

    A real-time lightweight object detection algorithm based on improved you only look once version 8 for unmanned surface vehicle

    Yinfeng Gong, Jiucai Jin, Deqing Liu, and Peng Ren. A real-time lightweight object detection algorithm based on improved you only look once version 8 for unmanned surface vehicle. Eng. Appl. Artif. Intell. , 152:110798, 2025

    2025

  16. [24]

    CGNet : A light-weight context guided network for semantic segmentation

    Tianyi Wu, Sheng Tang, Rui Zhang, Juan Cao, and Yongdong Zhang. CGNet : A light-weight context guided network for semantic segmentation. IEEE Trans. Image Process. , 30:1169--1179, 2020

    2020

  17. [25]

    Highly efficient salient object detection with 100k parameters

    Shang-Hua Gao, Yong-Qiang Tan, Ming-Ming Cheng, Chengze Lu, Yunpeng Chen, and Shuicheng Yan. Highly efficient salient object detection with 100k parameters. In Proc. Eur. Conf. Comput. Vis. , pages 702--721, 2020

    2020

  18. [26]

    A lightweight and high-accuracy framework for printed circuit board defect detection

    Yingqiang Hou and Xindong Zhang. A lightweight and high-accuracy framework for printed circuit board defect detection. Eng. Appl. Artif. Intell. , 148:110375, 2025

    2025

  19. [27]

    Bisenet v2: Bilateral network with guided aggregation for real-time semantic segmentation

    Changqian Yu, Changxin Gao, Jingbo Wang, Gang Yu, Chunhua Shen, and Nong Sang. Bisenet v2: Bilateral network with guided aggregation for real-time semantic segmentation. Int. J. Comput. Vis. , 129:3051--3068, 2021

    2021

  20. [28]

    TSD-DETR : A lightweight real-time detection transformer of traffic sign detection for long-range perception of autonomous driving

    Lili Zhang, Kang Yang, Yucheng Han, Jing Li, Wei Wei, Hongxin Tan, Pei Yu, Ke Zhang, and Xudong Yang. TSD-DETR : A lightweight real-time detection transformer of traffic sign detection for long-range perception of autonomous driving. Eng. Appl. Artif. Intell. , 139:109536, 2025

    2025

  21. [29]

    Real-time object detection network in uav-vision based on cnn and transformer

    Tao Ye, Wenyang Qin, Zongyang Zhao, Xiaozhi Gao, Xiangpeng Deng, and Yu Ouyang. Real-time object detection network in uav-vision based on cnn and transformer. IEEE Trans. Instrum. Meas. , 72:1--13, 2023

    2023

  22. [30]

    SegFormer : Simple and efficient design for semantic segmentation with transformers

    Enze Xie, Wenhai Wang, Zhiding Yu, Anima Anandkumar, Jose M Alvarez, and Ping Luo. SegFormer : Simple and efficient design for semantic segmentation with transformers. Advances in neural information processing systems , 34:12077--12090, 2021

    2021

  23. [31]

    Mobilevit: Light-weight, general-purpose, and mobile-friendly vision transformer

    Sachin Mehta and Mohammad Rastegari. Mobilevit: Light-weight, general-purpose, and mobile-friendly vision transformer. In Int. Conf. Learn. Represent. , pages 1--26, 2022

    2022

  24. [32]

    Efficientformer: Vision transformers at mobilenet speed

    Yanyu Li, Geng Yuan, Yang Wen, Ju Hu, Georgios Evangelidis, Sergey Tulyakov, Yanzhi Wang, and Jian Ren. Efficientformer: Vision transformers at mobilenet speed. Advances in Neural Information Processing Systems , 35:12934--12949, 2022

    2022

  25. [33]

    Fastvit: A fast hybrid vision transformer using structural reparameterization

    Pavan Kumar Anasosalu Vasu, James Gabriel, Jeff Zhu, Oncel Tuzel, and Anurag Ranjan. Fastvit: A fast hybrid vision transformer using structural reparameterization. In Proc. IEEE Int. Conf. Comput. Vis. , pages 5785--5795, 2023

    2023

  26. [34]

    Seaformer: Squeeze-enhanced axial transformer for mobile semantic segmentation

    Qiang Wan, Zilong Huang, Jiachen Lu, Gang Yu, and Li Zhang. Seaformer: Squeeze-enhanced axial transformer for mobile semantic segmentation. In Int. Conf. Learn. Represent. , pages 1--19, 2023

    2023

  27. [35]

    A survey on vision transformer

    Kai Han, Yunhe Wang, Hanting Chen, Xinghao Chen, Jianyuan Guo, Zhenhua Liu, Yehui Tang, An Xiao, Chunjing Xu, Yixing Xu, et al. A survey on vision transformer. IEEE Trans. Pattern Anal. Mach. Intell. , 45(1):87--110, 2022

    2022

  28. [36]

    Encoder-decoder with atrous separable convolution for semantic image segmentation

    Liang-Chieh Chen, Yukun Zhu, George Papandreou, Florian Schroff, and Hartwig Adam. Encoder-decoder with atrous separable convolution for semantic image segmentation. In Proc. Lect. Notes Comput. Sci. , pages 801--818, 2018

    2018

  29. [37]

    Scale-aware trident networks for object detection

    Yanghao Li, Yuntao Chen, Naiyan Wang, and Zhaoxiang Zhang. Scale-aware trident networks for object detection. In Proc. IEEE Int. Conf. Comput. Vis. , pages 6054--6063, 2019

    2019

  30. [38]

    Deep dual-resolution networks for real-time and accurate semantic segmentation of traffic scenes

    Huihui Pan, Yuanduo Hong, Weichao Sun, and Yisong Jia. Deep dual-resolution networks for real-time and accurate semantic segmentation of traffic scenes. IEEE Trans. Intell. Transp. Syst. , 24(3):3448--3460, 2022

    2022

  31. [39]

    Poolnet+: Exploring the potential of pooling for salient object detection

    Jiang-Jiang Liu, Qibin Hou, Zhi-Ang Liu, and Ming-Ming Cheng. Poolnet+: Exploring the potential of pooling for salient object detection. IEEE Trans. Pattern Anal. Mach. Intell. , 45(1):887--904, 2022

    2022

  32. [40]

    P2t: Pyramid pooling transformer for scene understanding

    Yu-Huan Wu, Yun Liu, Xin Zhan, and Ming-Ming Cheng. P2t: Pyramid pooling transformer for scene understanding. IEEE Trans. Pattern Anal. Mach. Intell. , 45(11):12760--12771, 2022

    2022

  33. [41]

    Pyramid vision transformer: A versatile backbone for dense prediction without convolutions

    Wenhai Wang, Enze Xie, Xiang Li, Deng-Ping Fan, Kaitao Song, Ding Liang, Tong Lu, Ping Luo, and Ling Shao. Pyramid vision transformer: A versatile backbone for dense prediction without convolutions. In Proc. IEEE Int. Conf. Comput. Vis. , pages 568--578, 2021

    2021

  34. [42]

    Hivit: Hierarchical attention-based transformer for multi-scale whole slide histopathological image classification

    Jinze Yu, Shuo Li, Luxin Tan, Haoyi Zhou, Zhongwu Li, and Jianxin Li. Hivit: Hierarchical attention-based transformer for multi-scale whole slide histopathological image classification. Expert. Sys. Appl. , 277:127164, 2025

    2025

  35. [43]

    Multiscale global context network for semantic segmentation of high-resolution remote sensing images

    Qiaolin Zeng, Jingxiang Zhou, Jinhua Tao, Liangfu Chen, Xuerui Niu, and Yumeng Zhang. Multiscale global context network for semantic segmentation of high-resolution remote sensing images. IEEE Trans. Geosci. Remote Sens. , 62:1--13, 2024

    2024

  36. [44]

    DUSA-UNet : Dual sparse attentive u-net for multiscale road network extraction

    Jie Song, Yue Sun, Ziyun Cai, Liang Xiao, Yawen Huang, and Yefeng Zheng. DUSA-UNet : Dual sparse attentive u-net for multiscale road network extraction. IEEE Trans. Geosci. Remote Sens. , 2025

    2025

  37. [45]

    View from the top: Hierarchies and reverse hierarchies in the visual system

    Shaul Hochstein and Merav Ahissar. View from the top: Hierarchies and reverse hierarchies in the visual system. Neuron , 36(5):791--804, 2002

    2002

  38. [46]

    The free-energy principle: a unified brain theory? Nat

    Karl Friston. The free-energy principle: a unified brain theory? Nat. Rev. Neurosci. , 11(2):127--138, 2010

    2010

  39. [47]

    Object perception as bayesian inference

    Daniel Kersten, Pascal Mamassian, and Alan Yuille. Object perception as bayesian inference. Annu. Rev. Psychol. , 55(1):271--304, 2004

    2004

  40. [48]

    Res2Net : A new multi-scale backbone architecture

    Shang-Hua Gao, Ming-Ming Cheng, Kai Zhao, Xin-Yu Zhang, Ming-Hsuan Yang, and Philip Torr. Res2Net : A new multi-scale backbone architecture. IEEE Trans. Pattern Anal. Mach. Intell. , 43(2):652--662, 2021

    2021

  41. [49]

    Deep high-resolution representation learning for visual recognition

    Jingdong Wang, Ke Sun, Tianheng Cheng, Borui Jiang, Chaorui Deng, Yang Zhao, Dong Liu, Yadong Mu, Mingkui Tan, Xinggang Wang, et al. Deep high-resolution representation learning for visual recognition. IEEE Trans. Pattern Anal. Mach. Intell. , 43(10):3349--3364, 2020

    2020

  42. [50]

    Cross-stage feature fusion and efficient self-attention for salient object detection

    Xiaofeng Xia and Yingdong Ma. Cross-stage feature fusion and efficient self-attention for salient object detection. J Visual Commun Image Represent. , 104:104271, 2024

    2024

  43. [51]

    Rewrite the stars

    Xu Ma, Xiyang Dai, Yue Bai, Yizhou Wang, and Yun Fu. Rewrite the stars. In Proc. IEEE Conf. Comput. Vis. Pattern Recognit. , pages 5694--5703, 2024

    2024

  44. [52]

    MCnet : Multiple context information segmentation network of no-service rail surface defects

    Defu Zhang, Kechen Song, Jing Xu, Yu He, Menghui Niu, and Yunhui Yan. MCnet : Multiple context information segmentation network of no-service rail surface defects. IEEE Trans. Instrum. Meas. , 70:1--9, 2020

    2020

  45. [53]

    A noise robust method based on completed local binary patterns for hot-rolled steel strip surface defects

    Kechen Song and Yunhui Yan. A noise robust method based on completed local binary patterns for hot-rolled steel strip surface defects. Appl. Surf. Sci. , 285:858--864, 2013

    2013

  46. [54]

    A simple pooling-based design for real-time salient object detection

    Jiang-Jiang Liu, Qibin Hou, Ming-Ming Cheng, Jiashi Feng, and Jianmin Jiang. A simple pooling-based design for real-time salient object detection. In Proc. IEEE Conf. Comput. Vis. Pattern Recognit. , pages 3917--3926, 2019

    2019

  47. [55]

    U ^ 2 -Net : Going deeper with nested u-structure for salient object detection

    Xuebin Qin, Zichen Zhang, Chenyang Huang, Masood Dehghan, Osmar R Zaiane, and Martin Jagersand. U ^ 2 -Net : Going deeper with nested u-structure for salient object detection. Pattern Recognit. , 106:107404, 2020

    2020

  48. [56]

    EDN : Salient object detection via extremely-downsampled network

    Yu-Huan Wu, Yun Liu, Le Zhang, Ming-Ming Cheng, and Bo Ren. EDN : Salient object detection via extremely-downsampled network. IEEE Trans. Image Process. , 31:3125--3136, 2022

    2022

  49. [57]

    Bisenet: Bilateral segmentation network for real-time semantic segmentation

    Changqian Yu, Jingbo Wang, Chao Peng, Changxin Gao, Gang Yu, and Nong Sang. Bisenet: Bilateral segmentation network for real-time semantic segmentation. In Proc. Eur. Conf. Comput. Vis. , pages 325--341, 2018

    2018

  50. [58]

    Lednet: A lightweight encoder-decoder network for real-time semantic segmentation

    Yu Wang, Quan Zhou, Jia Liu, Jian Xiong, Guangwei Gao, Xiaofu Wu, and Longin Jan Latecki. Lednet: A lightweight encoder-decoder network for real-time semantic segmentation. In Proc. IEEE Int. Conf. Image Process , pages 1860--1864, 2019

    2019

  51. [59]

    Lightweight salient object detection via hierarchical visual perception learning

    Yun Liu, Yu-Chao Gu, Xin-Yu Zhang, Weiwei Wang, and Ming-Ming Cheng. Lightweight salient object detection via hierarchical visual perception learning. IEEE Trans. Cybern. , 51(9):4439--4449, 2020

    2020

  52. [60]

    A convnet for the 2020s

    Zhuang Liu, Hanzi Mao, Chao-Yuan Wu, Christoph Feichtenhofer, Trevor Darrell, and Saining Xie. A convnet for the 2020s. In Proc. IEEE Conf. Comput. Vis. Pattern Recognit. , pages 11976--11986, 2022

    2022

Pith tools

Reviewed August 5, 2026 · model on record in the stance chip above.