Pith. sign in

REVIEW 3 major objections 2 minor 67 references

Interleaved Transceiver Design for a Continuous- Transmission MIMO-OFDM ISAC System

T0 review · 3 major / 2 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 This paper claims that an interleaved time-domain transceiver for continuous-transmission MIMO-OFDM can jointly shape one waveform for constructive-interference communication, an IMSR-based radar beampattern, interference-free continuous tr

desk verdict The submitted full text is a different paper (continual video instance segmentation), so the ISAC abstract cannot be evaluated; this is a submission error, not a paper. read the letter →

arxiv 2508.10430 v1 pith:KFTRLHQI submitted 2025-08-14 eess.SP

Yating Chen , Cai Wen , Yan Huang , Jinye Peng , Wei Hong , Timothy N. Davidson This is my paper
classification eess.SP
keywords integratedsensingandcommunicationMIMO-OFDMcontinuoustransmissioninterleavedtransceiverconstructiveinterferencemainlobe-to-sideloberatiopeak-to-averagepoweralternatingdirectionpenaltymethod
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 argues that in a continuously transmitting MIMO-OFDM integrated sensing and communication system, an interleaved transceiver can be designed entirely in the time domain so that one transmit waveform and one receive filter deliver both constructive-interference communication and an integrated-mainlobe-to-sidelobe-ratio radar beampattern, while suppressing inter-block interference and spurious peaks and bounding every sample's power. If true, the same hardware could sense and communicate without separate radar and communication waveforms, and the per-sample power constraints would keep the peak-to-average ratio manageable. The design problem is said to be solved by alternating optimization with successive convex approximation, and the inner alternating direction penalty method is claimed to be provably convergent, including the case of more than two auxiliary variables for the first time. A sympathetic reader would take the abstract at face value, but the supplied full text is an unrelated manuscript, so none of these statements can currently be checked against the body of the paper.

What carries the argument

The central object is the interleaved time-domain transmit waveform together with its designed receive filter; the radar beampattern quality is measured by the integrated mainlobe-to-sidelobe ratio (IMSR), and communication quality is enforced through constructive-interference (CI) constraints. The algorithmic machinery is an alternating-optimization (AO) outer loop, successive convex approximation (SCA) for the waveform subproblem, and an alternating direction penalty method as the inner accelerated solver. The claimed novelty is a convergence proof for ADPM with more than two auxiliary variables, which the authors say is needed to solve the multi-block waveform design efficiently.

What would settle it

Implement the proposed AO-SCA-ADPM algorithm on a small MIMO-OFDM configuration and check whether the returned waveform satisfies the per-sample PAPR bound at the claimed convergence tolerance; if the beampattern still contains spurious peaks or inter-block interference reappears, the central structural claim fails. Alternatively, exhibit a three-auxiliary-variable ADPM instance satisfying the stated assumptions that does not converge, which would refute the claimed first-time convergence result.

Watch

Extended reading notes

Core claim

The central claim is that co-designing the transmit waveform and receive filter in the time domain for an interleaved MIMO-OFDM ISAC architecture can simultaneously: (i) shape transmitted symbols toward constructive interference at the communication user, (ii) shape the radar beampattern through the integrated mainlobe-to-sidelobe ratio (IMSR), (iii) eliminate inter-block interference and spurious peaks that arise in continuous transmission, and (iv) keep every time-sample power within a prescribed PAPR budget. The proposed solution method is an alternating-optimization outer loop whose waveform subproblem is handled by successive convex approximation, accelerated by an alternating direction

Load-bearing premise

The load-bearing premise is the claimed first-time convergence proof for the alternating direction penalty method with more than two auxiliary variables, together with the assumption that the conditions of that proof are met by the interleaved ISAC design problem; the supplied text contains no statement of the theorem or its proof.

Editorial extensions

If this is right

  • A continuous-transmission MIMO-OFDM system could use one interleaved time-domain waveform for both radar and communication, avoiding separate waveforms and the block-edge artifacts that cause spurious peaks.
  • Per-sample power constraints put PAPR control directly into the design, so the waveform produced by the optimizer is already close to what a power amplifier can transmit without additional peak-reduction processing.
  • If the ADPM convergence result holds, the SCA subproblems can be solved faster, making the joint design loop practical for realistic array and symbol-block sizes.
  • Eliminating inter-block interference and spurious peaks by construction would make target detection more reliable in continuous transmission, which is the setting for joint sensing and communication systems such as automotive radar.
  • The authors report numerical simulations showing that the fast algorithm achieves comparable sensing and communication performance with greater computational efficiency than the slower exact subproblem solver.

Reading between the lines

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

  • Editorial note: the supplied full text is a different manuscript about continual video instance segmentation, so the ISAC claims and the ADPM convergence theorem exist for the reader only in the abstract; they should be treated as unverified until the actual derivation appears.
  • If the claimed ADPM theorem is valid, the same proof technique may transfer to other nonconvex multi-block optimization problems where penalty methods with more than two auxiliary variables were previously unproven.
  • A natural test beyond the paper's reported experiments is an ablation that fixes the ADPM iteration budget and measures how the PAPR versus IMSR trade-off degrades; this would expose whether the fast solver's practical convergence matches its theoretical guarantee.
  • A key check for practitioners is whether the assumptions of the convergence proof hold for the specific CI, IMSR, and per-sample power constraints of the ISAC problem; the abstract does not state those assumptions.
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 / 2 minor

Summary. The manuscript as submitted presents a title and abstract describing an interleaved transceiver design for a continuous-transmission MIMO-OFDM integrated sensing and communication (ISAC) system. The abstract claims an alternating optimization framework with successive convex approximation, an alternate direction penalty method (ADPM) inner solver, and establishes, for the first time, convergence of ADPM with more than two auxiliary variables. The claimed contributions include constructive-interference communication, integrated mainlobe-to-sidelobe ratio (IMSR) beampattern shaping, elimination of inter-block interference and spurious peaks, and per-sample PAPR constraints. However, the full text of the artifact is a completely different paper: 'CRISP: Contrastive Residual Injection and Semantic Prompting for Continual Video Instance Segmentation' (arXiv:2508.10432, cs.CV), with its own abstract, figures, tables, equations, and references. None of the ISAC formulation, ADPM algorithm, convergence theorem, or simulations appears anywhere in the body. The abstract functions as a claim without a supporting manuscript.

Significance. If the claimed ISAC results were actually presented and correct, the paper would be significant: a unified time-domain transceiver design achieving CI-based communication, IMSR-shaped sensing, inter-block-interference-free continuous transmission, per-sample PAPR control, and a provably convergent fast ADPM solver with more than two auxiliary variables would be a meaningful advance for MIMO-OFDM ISAC. However, no such content is present in the artifact. The body is an unrelated continual video instance segmentation paper with its own title, authors, and contributions. The central mathematical claim—first-time convergence of ADPM with more than two auxiliary variables—has no theorem statement, no assumptions, and no proof in the manuscript. No simulations for the ISAC system are reported. The artifact is internally inconsistent: the abstract and the body describe two unrelated works. Consequently, not a single claimed contribution can be evaluated or verified.

major comments (3)
  1. [Title/Abstract vs. Full Text] The title and abstract describe a MIMO-OFDM ISAC transceiver design, but the full text is 'CRISP: Contrastive Residual Injection and Semantic Prompting for Continual Video Instance Segmentation', arXiv:2508.10432 [cs.CV] by different authors. The body contains no ISAC system model, no optimization problem, no algorithm statement, no convergence theorem, and no ISAC simulations. The internal inconsistency is decisive: the manuscript submitted for review is not the manuscript described in its abstract.
  2. [Abstract, ADPM convergence claim] The abstract states that 'the convergence of ADPM is established, with convergence of the case of more than two auxiliary variables being established for the first time.' This is the load-bearing mathematical contribution supporting the computational-efficiency claim. The artifact contains no statement of this theorem, no assumptions (e.g., convexity, smoothness, penalty parameter conditions), and no proof. There is also no description of the ADPM algorithm or its auxiliary variables in the context of the ISAC problem. The claim is therefore entirely unsupported.
  3. [Entire ISAC contribution] Every substantive ISAC claim—constructive-interference communication, IMSR-based beampattern with no inter-block interference or spurious peaks, per-sample PAPR control, AO-SCA solution, and numerical validation—is absent from the body. No equations, algorithm pseudocode, or simulation figures/tables for the ISAC system appear. This is not a matter of a missing proof or a local gap; the entire contribution is missing, making the manuscript's central claim unverifiable.
minor comments (2)
  1. [Title page] The title page identifies the paper as arXiv:2508.10432v1, which is inconsistent with the assigned manuscript identifier 2508.10430. This reinforces the apparent mismatch between the submitted PDF and the intended paper.
  2. [Abstract] The abstract contains no references to prior ADPM or ISAC work, and it is not mirrored by any section headings or content in the body. Section headings in the body (Introduction, Continual Video Instance Segmentation, Experiments, Conclusion, etc.) relate exclusively to CRISP.

Circularity Check

0 steps flagged · score 0.0 of 10

No circularity can be identified: the supplied full text is a different paper (CRISP, arXiv:2508.10432) and contains none of the abstract's ISAC/ADPM derivation, so there is no reduction-by-construction to exhibit.

full rationale

The abstract describes an interleaved transceiver design for MIMO-OFDM ISAC, claiming an ADPM convergence result 'established for the first time.' The full text provided, however, is CRISP, a continual video instance segmentation paper (arXiv:2508.10432v1 [cs.CV]) by different authors, with its own abstract, figures, tables, and references. None of the ISAC formulation, the IMSR beampattern design, the inter-block interference analysis, the PAPR constraints, the AO-SCA framework, the ADPM algorithm, or the claimed convergence theorem appears anywhere in the document. Because the claimed derivation chain is entirely absent, there is no equation or fitted parameter that can be quoted to show that a prediction is equivalent to its input by construction, and there is no load-bearing self-citation chain to evaluate. The document mismatch and missing proof are serious completeness/integrity problems, but they are not circularity under the definitions used here. Accordingly, the honest circularity finding is 0, with no circular steps identified; the abstract's claims are unsupported within this artifact, not circularly derived.

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

The central ISAC claim rests on standard optimization machinery (AO, SCA, ADPM) and on domain modeling choices stated only at the abstract level: the OFDM/ISAC signal model, the CI detection model, and the IMSR beampattern criterion. No parameter values, masks, or penalty schedules are given in the artifact. No invented physical entities are introduced.

free parameters (4)
  • per-sample power constraint bound (PAPR limit)
    The abstract constrains the power of each time sample to manage PAPR, but the bound value is not stated; the design's feasible region depends on it.
  • IMSR beampattern weights (mainlobe/sidelobe mask)
    The integrated mainlobe-to-sidelobe ratio requires a chosen angular region weighting; the abstract does not specify it and the design output depends on it.
  • CI constraint margin
    Constructive-interference communication requires a per-symbol margin parameter; this is a tuning knob on which the feasibility of the optimization depends.
  • ADPM penalty parameter
    Alternating direction penalty methods converge depending on the penalty sequence; the abstract does not state the update rule or initial penalty.
assumptions (4)
  • domain assumption MIMO-OFDM ISAC signal model with interleaved time-domain transmission
    The abstract postulates this architecture; all sensing and communication metrics are defined relative to it.
  • domain assumption Constructive interference provides a valid formulation of the communication constraint
    CI-based design presumes symbol-level knowledge and a detection model that makes the constraint set convex; this is not stated or derived in the artifact.
  • domain assumption Integrated mainlobe-to-sidelobe ratio is an adequate radar directivity metric
    The abstract selects IMSR to ensure desirable directivity; this modeling choice supports the optimization objective and any validation.
  • standard math Convergence assumptions for AO and SCA
    AO and SCA convergence relies on standard assumptions such as differentiability and convexity of subproblems; invoked implicitly, no theorem statement is available.

how reviews work

0 comments
Cite this review

Pith. "Pith review of Interleaved Transceiver Design for a Continuous- Transmission MIMO-OFDM ISAC System." pith.science (2026). https://pith.science/paper/KFTRLHQI

@misc{pith2026250810430,
  author       = {Pith},
  title        = {Pith review of: Interleaved Transceiver Design for a Continuous- Transmission MIMO-OFDM ISAC System},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/KFTRLHQI}},
  note         = {Machine review of arXiv:2508.10430}
}
read the original abstract

This paper proposes an interleaved transceiver design method for a multiple-input multiple-output (MIMO) integrated sensing and communication (ISAC) system utilizing orthogonal frequency division multiplexing (OFDM) waveforms. We consider a continuous transmission system and focus on the design of the transmission signal and a receiving filter in the time domain for an interleaved transmission architecture. For communication performance, constructive interference (CI) is integrated into the optimization problem. For radar sensing performance, the integrated mainlobe-to-sidelobe ratio (IMSR) of the beampattern is considered to ensure desirable directivity. Additionally, we tackle the challenges of inter-block interference and eliminate the spurious peaks, which are crucial for accurate target detection. Regarding the hardware implementation aspect, the power of each time sample is constrained to manage the peak-to-average power ratio (PAPR). The design problem is addressed using an alternating optimization (AO) framework, with the subproblem for transmitted waveform design being solved via the successive convex approximation (SCA) method. To further enhance computational efficiency, the alternate direction penalty method (ADPM) is employed to solve the subproblems within the SCA iterations. The convergence of ADPM is established, with convergence of the case of more than two auxiliary variables being established for the first time. Numerical simulations validate the effectiveness of our transceiver design in achieving desirable performance in both radar sensing and communication, with the fast algorithm achieving comparable performance with greater computational efficiency.

Discussion (0). Sign in to comment.

Reference graph

Works this paper leans on

67 extracted references · 51 canonical work pages

  1. [1]

    instance-wise

    and [2] are motorbikes. As the video progresses, we observe that ECLIPSE mistakenly merges instances [1] and [2], which are two motorbikes separated by some distance on the road, into a single instance. We term this type of error “instance-wise” confusion. In the second sequence, instances

    work page

  2. [3]

    category-wise

    and [2] are sedans, and instance [1] is a hand. We observe that ECLIPSE sometimes misclassifies instance [2] as a motorbike. We define this type of error, where instances of the same video and category are misclassified into different categories, as “category-wise” confusion. In the third se- quence, the scene contains “person” category from the old task ...

    work page arXiv 2019

  3. [4]

    Backdoor attacks-resilient aggregation based on robust filtering of outliers in federated learning for image classification,

    N. Rodr ´ıguez-Barroso, E. Mart ´ınez-C´amara, M. V . Luz´on, and F. Her- rera, “Backdoor attacks-resilient aggregation based on robust filtering of outliers in federated learning for image classification,”Knowledge-Based Systems, vol. 245, p. 108588, 2022

    work page 2022

  4. [5]

    Multi-evidence filtering and fusion for multi-label classification, object detection and semantic segmentation based on weakly supervised learning,

    W. Ge, S. Yang, and Y . Yu, “Multi-evidence filtering and fusion for multi-label classification, object detection and semantic segmentation based on weakly supervised learning,” inProceedings of the IEEE conference on computer vision and pattern recognition, 2018, pp. 1277– 1286

    work page 2018

  5. [6]

    Structured attention network for referring image segmentation,

    L. Lin, P. Yan, X. Xu, S. Yang, K. Zeng, and G. Li, “Structured attention network for referring image segmentation,”IEEE Transactions on Multimedia, vol. 24, pp. 1922–1932, 2021

    work page 1922

  6. [7]

    Adaptive dual-axis style-based recalibration network with class-wise statistics loss for imbalanced medical image classification,

    X. Zhang, Z. Xiao, J. Ma, X. Wu, J. Zhao, S. Zhang, R. Li, Y . Pan, and J. Liu, “Adaptive dual-axis style-based recalibration network with class-wise statistics loss for imbalanced medical image classification,” IEEE Transactions on Image Processing, vol. 34, pp. 2081–2096, 2025

    work page

  7. [8]

    Searching dis- criminative regions for convolutional neural networks in fundus image classification with genetic algorithms,

    Y . Rong, T. Lin, H. Chen, Z. Fan, and X. Chen, “Searching dis- criminative regions for convolutional neural networks in fundus image classification with genetic algorithms,”IEEE Transactions on Image Processing, vol. 33, pp. 5949–5958, 2024

    work page 2024

  8. [9]

    Unsupervised person re-identification based on adaptive information supplementation and foreground enhancement,

    Q. Wang, Z. Huang, H. Fan, S. Fu, and Y . Tang, “Unsupervised person re-identification based on adaptive information supplementation and foreground enhancement,”IET Image Processing, vol. 18, no. 14, pp. 4680–4694, 2024

    work page 2024

Show all 67 references
  1. [10]

    Learning self- and cross-triplet context clues for human-object interaction detec- tion,

    W. Ren, J. Luo, W. Jiang, L. Qu, Z. Han, J. Tian, and H. Liu, “Learning self- and cross-triplet context clues for human-object interaction detec- tion,”IEEE Transactions on Circuits and Systems for Video Technology, vol. 34, no. 10, pp. 9760–9773, 2024

    2024

  2. [11]

    Com- monality autoencoder: Learning common features for change detection from heterogeneous images,

    Y . Wu, J. Li, Y . Yuan, A. K. Qin, Q.-G. Miao, and M.-G. Gong, “Com- monality autoencoder: Learning common features for change detection from heterogeneous images,”IEEE transactions on neural networks and learning systems, vol. 33, no. 9, pp. 4257–4270, 2021

    2021

  3. [12]

    Lightweight class incremental semantic segmentation without catastrophic forgetting,

    W. Cong, Y . Cong, and Y . Ren, “Lightweight class incremental semantic segmentation without catastrophic forgetting,”IEEE Transactions on Image Processing, pp. 1–1, 2025

    2025

  4. [13]

    Continual segmentation with disentangled objectness learning and class recognition,

    Y . Gong, S. Yu, X. Wang, and J. Xiao, “Continual segmentation with disentangled objectness learning and class recognition,” inProceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recogni- tion, 2024, pp. 3848–3857

    2024

  5. [14]

    Repre- sentation compensation networks for continual semantic segmentation,

    C.-B. Zhang, J.-W. Xiao, X. Liu, Y .-C. Chen, and M.-M. Cheng, “Repre- sentation compensation networks for continual semantic segmentation,” inProceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition, 2022, pp. 7053–7064

    2022

  6. [15]

    Deeplab: Semantic image segmentation with deep convolutional nets, atrous convolution, and fully connected crfs,

    L.-C. Chen, G. Papandreou, I. Kokkinos, K. Murphy, and A. L. Yuille, “Deeplab: Semantic image segmentation with deep convolutional nets, atrous convolution, and fully connected crfs,”IEEE transactions on pattern analysis and machine intelligence, vol. 40, no. 4, pp. 834–848, 2017

    2017

  7. [16]

    Segmenter: Trans- former for semantic segmentation,

    R. Strudel, R. Garcia, I. Laptev, and C. Schmid, “Segmenter: Trans- former for semantic segmentation,” inProceedings of the IEEE/CVF international conference on computer vision, 2021, pp. 7262–7272

    2021

  8. [17]

    Instance-aware semantic segmentation via multi-task network cascades,

    J. Dai, K. He, and J. Sun, “Instance-aware semantic segmentation via multi-task network cascades,” inProceedings of the IEEE conference on computer vision and pattern recognition, 2016, pp. 3150–3158

    2016

  9. [18]

    Fully convolutional instance- aware semantic segmentation,

    Y . Li, H. Qi, J. Dai, X. Ji, and Y . Wei, “Fully convolutional instance- aware semantic segmentation,” inProceedings of the IEEE conference on computer vision and pattern recognition, 2017, pp. 2359–2367

    2017

  10. [19]

    Solo: Segmenting objects by locations,

    X. Wang, T. Kong, C. Shen, Y . Jiang, and L. Li, “Solo: Segmenting objects by locations,” inEuropean conference on computer vision. Springer, 2020, pp. 649–665

    2020

  11. [20]

    Masked-attention mask transformer for universal image segmentation,

    B. Cheng, I. Misra, A. G. Schwing, A. Kirillov, and R. Girdhar, “Masked-attention mask transformer for universal image segmentation,” inProceedings of the IEEE/CVF conference on computer vision and pattern recognition, 2022, pp. 1290–1299

    2022

  12. [21]

    Mask2former for video instance segmentation,

    B. Cheng, A. Choudhuri, I. Misra, A. Kirillov, R. Girdhar, and A. G. Schwing, “Mask2former for video instance segmentation,” 2021. [Online]. Available: https://arxiv.org/abs/2112.10764

    2021 arXiv

  13. [22]

    Catastrophic forgetting in connectionist networks,

    R. M. French, “Catastrophic forgetting in connectionist networks,” Trends in cognitive sciences, vol. 3, no. 4, pp. 128–135, 1999

    1999

  14. [23]

    Catastrophic forgetting, rehearsal and pseudorehearsal,

    A. Robins, “Catastrophic forgetting, rehearsal and pseudorehearsal,” Connection Science, vol. 7, no. 2, pp. 123–146, 1995

    1995

  15. [24]

    Lifelong learning algorithms,

    S. Thrun, “Lifelong learning algorithms,” inLearning to learn. Springer, 1998, pp. 181–209

    1998

  16. [25]

    A continual learning survey: Defying forgetting in classification tasks,

    M. De Lange, R. Aljundi, M. Masana, S. Parisot, X. Jia, A. Leonardis, G. Slabaugh, and T. Tuytelaars, “A continual learning survey: Defying forgetting in classification tasks,”IEEE transactions on pattern analysis and machine intelligence, vol. 44, no. 7, pp. 3366–3385, 2021

    2021

  17. [26]

    Calibrating cnns for lifelong learning,

    P. Singh, V . K. Verma, P. Mazumder, L. Carin, and P. Rai, “Calibrating cnns for lifelong learning,”Advances in Neural Information Processing Systems, vol. 33, pp. 15 579–15 590, 2020

    2020

  18. [27]

    Federated class-incremental learning,

    J. Dong, L. Wang, Z. Fang, G. Sun, S. Xu, X. Wang, and Q. Zhu, “Federated class-incremental learning,” inIEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR), June 2022

    2022

  19. [28]

    No one left behind: Real-world federated class-incremental learning,

    J. Dong, H. Li, Y . Cong, G. Sun, Y . Zhang, and L. Van Gool, “No one left behind: Real-world federated class-incremental learning,”IEEE Transactions on Pattern Analysis and Machine Intelligence, vol. 46, no. 4, pp. 2054–2070, 2024

  20. [29]

    Task-specific normalization for continual learning of blind image quality models,

    W. Zhang, K. Ma, G. Zhai, and X. Yang, “Task-specific normalization for continual learning of blind image quality models,”IEEE Transactions on Image Processing, vol. 33, pp. 1898–1910, 2024

    1910

  21. [30]

    car” class in a new “bus

    allow a deep neural network to adapt to new tasks while preserving previously learned knowledge. In continual image segmentation tasks, many methods have made significant progress. Knowledge distillation based continual segmentation methods [13], [31], [32] alleviate catastrop...

    2025 arXiv

  22. [31]

    Multi- scale feature alignment for continual learning of unlabeled domains,

    K. Thandiackal, L. Piccinelli, R. Gupta, P. Pati, and O. Goksel, “Multi- scale feature alignment for continual learning of unlabeled domains,” IEEE Transactions on Medical Imaging, vol. 43, no. 7, pp. 2599–2609, 2024

    2024

  23. [32]

    Der: Dynamically expandable representation for class incremental learning,

    S. Yan, J. Xie, and X. He, “Der: Dynamically expandable representation for class incremental learning,” inProceedings of the IEEE/CVF confer- ence on computer vision and pattern recognition, 2021, pp. 3014–3023

    2021

  24. [33]

    Modeling the background for incremental learning in semantic segmentation,

    F. Cermelli, M. Mancini, S. R. Bulo, E. Ricci, and B. Caputo, “Modeling the background for incremental learning in semantic segmentation,” in Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition, 2020, pp. 9233–9242

    2020

  25. [34]

    Plop: Learning without forgetting for continual semantic segmentation,

    A. Douillard, Y . Chen, A. Dapogny, and M. Cord, “Plop: Learning without forgetting for continual semantic segmentation,” inProceedings of the IEEE/CVF conference on computer vision and pattern recognition, 2021, pp. 4040–4050

    2021

  26. [35]

    Eclipse: Efficient continual learning in panoptic segmentation with visual prompt tuning,

    B. Kim, J. Yu, and S. J. Hwang, “Eclipse: Efficient continual learning in panoptic segmentation with visual prompt tuning,” inProceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition, 2024, pp. 3346–3356

    2024

  27. [36]

    Promptfusion: Decoupling stability and plasticity for continual learning,

    H. Chen, Z. Wu, X. Han, M. Jia, and Y .-G. Jiang, “Promptfusion: Decoupling stability and plasticity for continual learning,” inEuropean Conference on Computer Vision. Springer, 2024, pp. 196–212

    2024

  28. [37]

    Coda-prompt: Contin- ual decomposed attention-based prompting for rehearsal-free continual learning,

    J. S. Smith, L. Karlinsky, V . Gutta, P. Cascante-Bonilla, D. Kim, A. Arbelle, R. Panda, R. Feris, and Z. Kira, “Coda-prompt: Contin- ual decomposed attention-based prompting for rehearsal-free continual learning,” inProceedings of the IEEE/CVF conference on computer vision an...

    2023

  29. [38]

    Attriclip: A non-incremental learner for incremental knowledge learn- ing,

    R. Wang, X. Duan, G. Kang, J. Liu, S. Lin, S. Xu, J. L ¨u, and B. Zhang, “Attriclip: A non-incremental learner for incremental knowledge learn- ing,” inProceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition, 2023, pp. 3654–3663

    2023

  30. [39]

    Dualprompt: Complementary prompting for rehearsal-free continual learning,

    Z. Wang, Z. Zhang, S. Ebrahimi, R. Sun, H. Zhang, C.-Y . Lee, X. Ren, G. Su, V . Perot, J. Dyet al., “Dualprompt: Complementary prompting for rehearsal-free continual learning,” inEuropean conference on computer vision. Springer, 2022, pp. 631–648

    2022

  31. [40]

    Learning to prompt for continual learning,

    Z. Wang, Z. Zhang, C.-Y . Lee, H. Zhang, R. Sun, X. Ren, G. Su, V . Perot, J. Dy, and T. Pfister, “Learning to prompt for continual learning,” inProceedings of the IEEE/CVF conference on computer vision and pattern recognition, 2022, pp. 139–149

    2022

  32. [41]

    Semantic residual prompts for continual learning,

    M. Menabue, E. Frascaroli, M. Boschini, E. Sangineto, L. Bonicelli, A. Porrello, and S. Calderara, “Semantic residual prompts for continual learning,” inEuropean Conference on Computer Vision. Springer, 2024, pp. 1–18

    2024

  33. [42]

    Ssul: Semantic segmentation with unknown label for exemplar-based class-incremental learning,

    S. Cha, Y . Yoo, T. Moonet al., “Ssul: Semantic segmentation with unknown label for exemplar-based class-incremental learning,”Advances in neural information processing systems, vol. 34, pp. 10 919–10 930, 2021

    2021

  34. [43]

    Class-incremental learning for action recognition in videos,

    J. Park, M. Kang, and B. Han, “Class-incremental learning for action recognition in videos,” inProceedings of the IEEE/CVF international conference on computer vision, 2021, pp. 13 698–13 707

    2021

  35. [44]

    vclimb: A novel video class incremental learning bench- mark,

    A. Villa, K. Alhamoud, V . Escorcia, F. Caba, J. L. Alc ´azar, and B. Ghanem, “vclimb: A novel video class incremental learning bench- mark,” inProceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition, 2022, pp. 19 035–19 044

    2022

  36. [45]

    Space-time prompting for video class-incremental learning,

    Y . Pei, Z. Qing, S. Zhang, X. Wang, Y . Zhang, D. Zhao, and X. Qian, “Space-time prompting for video class-incremental learning,” inPro- ceedings of the IEEE/CVF International Conference on Computer Vi- sion, 2023, pp. 11 932–11 942

    2023

  37. [46]

    Stsp: Spatial-temporal subspace projection for video class-incremental learning,

    H. Cheng, S. Yang, C. Wang, J. T. Zhou, A. C. Kot, and B. Wen, “Stsp: Spatial-temporal subspace projection for video class-incremental learning,” inEuropean Conference on Computer Vision. Springer, 2024, pp. 374–391

    2024

  38. [47]

    Csta: Spatial-temporal causal adaptive learning for exemplar-free video class- 13 incremental learning,

    T. Chen, H. Liu, C. H. Lim, J. See, X. Gao, J. Hou, and W. Lin, “Csta: Spatial-temporal causal adaptive learning for exemplar-free video class- 13 incremental learning,”IEEE Transactions on Circuits and Systems for Video Technology, 2025

    2025

  39. [48]

    When video classification meets incremental classes,

    H. Zhao, X. Qin, S. Su, Y . Fu, Z. Lin, and X. Li, “When video classification meets incremental classes,” inProceedings of the 29th ACM International Conference on Multimedia, 2021, pp. 880–889

    2021

  40. [49]

    Visualizing data using t-sne,

    L. v. d. Maaten and G. Hinton, “Visualizing data using t-sne,”Journal of machine learning research, vol. 9, no. Nov, pp. 2579–2605, 2008

    2008

  41. [50]

    Principal component analysis,

    H. Abdi and L. J. Williams, “Principal component analysis,”Wiley interdisciplinary reviews: computational statistics, vol. 2, no. 4, pp. 433– 459, 2010

    2010

  42. [51]

    Principal component analysis,

    I. Jolliffe, “Principal component analysis,” inInternational encyclopedia of statistical science. Springer, 2011, pp. 1094–1096

    2011

  43. [52]

    Video instance segmentation,

    L. Yang, Y . Fan, and N. Xu, “Video instance segmentation,” inInterna- tional Journal of Computer Vision,, 2019

    2019

  44. [53]

    The 3rd large-scale video object segmentation challenge - video instance segmentation track,

    L. Yang, Y . Fan, Y . Fu, and N. Xu, “The 3rd large-scale video object segmentation challenge - video instance segmentation track,” Jun. 2021

    2021

  45. [54]

    Mask r-cnn,

    K. He, G. Gkioxari, P. Doll ´ar, and R. Girshick, “Mask r-cnn,” in Proceedings of the IEEE international conference on computer vision, 2017, pp. 2961–2969

    2017

  46. [55]

    End-to-end object detection with transformers,

    N. Carion, F. Massa, G. Synnaeve, N. Usunier, A. Kirillov, and S. Zagoruyko, “End-to-end object detection with transformers,” in European conference on computer vision. Springer, 2020, pp. 213– 229

    2020

  47. [56]

    End-to-end video instance segmentation with transformers,

    Y . Wang, Z. Xu, X. Wang, C. Shen, B. Cheng, H. Shen, and H. Xia, “End-to-end video instance segmentation with transformers,” inPro- ceedings of the IEEE/CVF conference on computer vision and pattern recognition, 2021, pp. 8741–8750

    2021

  48. [57]

    Video instance segmenta- tion using inter-frame communication transformers,

    S. Hwang, M. Heo, S. W. Oh, and S. J. Kim, “Video instance segmenta- tion using inter-frame communication transformers,”Advances in Neural Information Processing Systems, vol. 34, pp. 13 352–13 363, 2021

    2021

  49. [58]

    Class-incremental instance segmentation via multi-teacher networks,

    Y . Gu, C. Deng, and K. Wei, “Class-incremental instance segmentation via multi-teacher networks,” inProceedings of the AAAI Conference on Artificial Intelligence, vol. 35, no. 2, 2021, pp. 1478–1486

    2021

  50. [59]

    Rbc: Rectifying the biased context in continual semantic segmentation,

    H. Zhao, F. Yang, X. Fu, and X. Li, “Rbc: Rectifying the biased context in continual semantic segmentation,” inEuropean Conference on Computer Vision. Springer, 2022, pp. 55–72

    2022

  51. [60]

    Class similarity weighted knowledge distillation for continual semantic seg- mentation,

    M. H. Phan, S. L. Phung, L. Tran-Thanh, A. Bouzerdoumet al., “Class similarity weighted knowledge distillation for continual semantic seg- mentation,” inProceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition, 2022, pp. 16 866–16 875

    2022

  52. [61]

    Deformable detr: Deformable transformers for end-to-end object detection,

    X. Zhu, W. Su, L. Lu, B. Li, X. Wang, and J. Dai, “Deformable detr: Deformable transformers for end-to-end object detection,”arXiv preprint arXiv:2010.04159, 2020

    2010 arXiv

  53. [62]

    Comformer: Continual learn- ing in semantic and panoptic segmentation,

    F. Cermelli, M. Cord, and A. Douillard, “Comformer: Continual learn- ing in semantic and panoptic segmentation,” inProceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition, 2023, pp. 3010–3020

    2023

  54. [63]

    Strike a balance in continual panoptic segmentation,

    J. Chen, R. Cong, Y . Luo, H. H. S. Ip, and S. Kwong, “Strike a balance in continual panoptic segmentation,” inEuropean Conference on Computer Vision. Springer, 2024, pp. 126–142

    2024

  55. [64]

    Learning to prompt for vision- language models,

    K. Zhou, J. Yang, C. C. Loy, and Z. Liu, “Learning to prompt for vision- language models,”International Journal of Computer Vision, vol. 130, no. 9, pp. 2337–2348, 2022

    2022

  56. [65]

    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. Clarket al., “Learning transferable visual models from natural language supervision,” inInternational conference on machine learning. PmLR, 2021, pp. 8748–8763

    2021

  57. [66]

    Context-aware video instance segmentation,

    S. Lee, J. Seo, K. Han, M. Choi, and S. Im, “Context-aware video instance segmentation,”arXiv preprint arXiv:2407.03010, 2024

    2024 arXiv

  58. [67]

    Modeling the background for incremental learning in semantic seg- mentation,

    F. Cermelli, M. Mancini, S. Rota Bul `o, E. Ricci, and B. Caputo, “Modeling the background for incremental learning in semantic seg- mentation,” in2020 IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR), 2020, pp. 9230–9239

    2020

  59. [68]

    Combo: Conflict mitigation via branched optimization for class incremental segmentation,

    K. Fang, A. Zhang, G. Gao, J. Jiao, C. H. Liu, and Y . Wei, “Combo: Conflict mitigation via branched optimization for class incremental segmentation,”arXiv preprint arXiv:2504.04156, 2025

    2025 arXiv

Pith tools

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