Pith. sign in

REVIEW 4 major objections 5 minor 1 cited by

Can We Get Rid of Handcrafted Feature Extractors? SparseViT: Nonsemantics-Centered, Parameter-Efficient Image Manipulation Localization through Spare-Coding Transformer

T0 review · 4 major / 5 minor · reviewed 2026-08-11 · deepseek-v4-flash

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

Pith's one-line read SparseViT shows that sparse self-attention can replace handcrafted feature extractors in image manipulation localization, reporting the strongest average F1 and AUC across four benchmarks while cutting FLOPs by up to 80%.

desk verdict SparseViT makes a credible case that IML can drop handcrafted extractors, but the headline generalization claim is weakened by tuning the sparsity schedule and fusion head on the test sets. read the letter →

arxiv 2412.14598 v2 pith:ZWYJUMOZ submitted 2024-12-19 cs.CV

Lei Su , Xiaochen Ma , Xuekang Zhu , Chaoqun Niu , Zeyu Lei , Ji-Zhe Zhou This is my paper · ORCID
classification cs.CV
keywords imagemanipulationlocalizationsparseself-attentionvisiontransformernon-semanticfeaturessemantic-agnostichandcraftedfeatureextractorslearnablefusionforensics
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 sets out to answer a pointed question: can image manipulation localization work without handcrafted feature extractors, which are normally used to supply the non-semantic clues (noise, frequency, compression artifacts) that reveal tampering? Its answer is yes, via a sparse-coding vision transformer called SparseViT. The core idea is that non-semantic features are locally independent and consistent across image patches unless manipulation occurs, so sparse interactions between patches suffice to capture them, while dense attention is what drags a model toward semantics. The paper shows that replacing global self-attention with sparse self-attention and fusing multi-scale outputs with a lightweight learnable head yields the best average pixel F1 and AUC across COVERAGE, Columbia, CASIAv1, and NIST16 without any handcrafted extractor, at a fraction of the FLOPs of prior methods. A sympathetic reader would see this as evidence that the handcrafted-extractor design pattern can be retired in favor of an adaptive, learned mechanism.

What carries the argument

The load-bearing mechanism is sparse self-attention with a sparsity rate $S$, which reshapes a feature map of shape $H \times W \times C$ into $S \times S$ non-overlapping blocks of shape $H/S \times W/S$ and restricts token-to-token attention to within each block. This removes long-range key-value pairs and, the paper argues, suppresses semantic expression because semantics require dense, continuous interactions across patches. The exponential sparsity schedule across stages yields multi-scale features that range from more semantic (low $S$) to more non-semantic (high $S$). A second component, the Learnable Feature Fusion (LFF) head, multiplies each of the six feature maps by a learnable scalar $\gamma$ initialized near $10^{-6}$ before summing and reducing to a single-channel prediction mask, adapting the fusion weights per image.

What would settle it

Measure the per-patch distribution of high-frequency noise or compression residual statistics on a set of pristine images: if these statistics vary substantially from patch to patch within a single image, the consistency premise fails. Or test SparseViT on a manipulation that changes global statistics without creating local inconsistencies, such as a uniform color cast over the whole image; under the paper's mechanism, performance should collapse if the model can only exploit locally independent non-semantic features.

Watch

Extended reading notes

Core claim

On the paper's own terms, SparseViT's discovery is that sparsity in self-attention is not merely a computational optimization but a representational intervention: it breaks the dense, global interactions that carry semantic meaning and thereby forces the model to attend to manipulation-sensitive local artifacts. The authors argue that semantic features vary drastically across patches and require continuous, long-range attention to assemble, whereas non-semantic features such as noise and high-frequency inconsistencies are consistent across non-manipulated patches and therefore can be captured by discrete, local interactions. SparseViT implements this by dividing each feature map into $S \times S$ non-overlapping blocks and running self-attention only within each block, with an exponentially decreasing sparsity rate across stages; the resulting multi-scale features are combined by a learnable fusion head with small per-feature scaling parameters. In cross-dataset experiments where all models are trained on the CAT-Net dataset, SparseViT reports the highest average pixel-level F1 (0.671) and AUC (0.937) on the four standard benchmarks, with 46.2 GFLOPs, the lowest among compared methods.

Load-bearing premise

The architecture rests on the premise that non-semantic manipulation traces are approximately uniform across non-manipulated patches and only deviate where manipulation occurred, so sparse local attention is sufficient to capture them; the paper does not measure this patch independence directly.

Editorial extensions

If this is right

  • Handcrafted extractors such as BayarConv, SRM, DCT, and Noiseprint can be removed from an IML pipeline: replacing them with sparse self-attention improves average pixel F1 from 0.627 (TruFor) to 0.671 and average AUC from 0.895 to 0.937 on the four shared benchmarks.
  • The sparsity mechanism cuts computation sharply: SparseViT runs at 46.2 GFLOPs on 512×512 input, about 80% less than TruFor's 236.5 GFLOPs, while using 50.3M parameters.
  • Sparse attention is the causal contributor, not just a cheaper approximation: ablation shows it improves average F1 over global attention on all five datasets, whereas plugging Bayar, Sobel, DCT, or SRM into the same backbone sometimes degrades performance.
  • The sparsification recipe transfers across transformer families: applying it to vanilla ViT and VOLO improves their average F1, suggesting the benefit is not specific to the Uniformer-like backbone.
  • A learnable fusion head (LFF) outperforms single-scale supervision and an MLP head, and does so with 0.66M parameters and 0.68 GFLOPs, much lighter than ASPP or AFF alternatives.

Reading between the lines

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

  • If the patch-independence premise holds, the same sparse-attention trick should transfer to other forensic dense-prediction tasks that rely on local inconsistencies, such as deepfake video segmentation or steganalysis; a direct test would be to plug the sparsified head into those task backbones.
  • The paper's fixed-threshold F1 reporting is conservative, but its AUC gains suggest that threshold choice could shift rankings; a practical extension is a calibration study reporting F1 at the optimal F1-weighted threshold across all baselines.
  • The architecture still uses ImageNet-pretrained Uniformer weights, so the claim of 'adaptive extraction' is about the attention pattern, not about learning from scratch; an inference is that a randomly initialized SparseViT trained on forensic data alone would isolate how much of the benefit comes from sparsity versus pretraining.
Share X Bluesky LinkedIn Reddit HN

Signed reviews

No signed human review yet.

Editorial analysis

A structured set of objections, weighed in public.

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

Referee Report

4 major / 5 minor

Summary. The paper proposes SparseViT, a vision transformer for image manipulation localization that replaces dense global self-attention with blockwise sparse self-attention at increasing sparsity rates across stages, and pairs this with a lightweight learnable feature-fusion head (LFF). The authors argue that semantic features require dense global interactions whereas non-semantic manipulation traces are locally independent, so sparse attention suppresses semantics and adaptively extracts non-semantic features without handcrafted forensic extractors. The model is trained on the CAT-Net dataset and evaluated on COVERAGE, Columbia, CASIAv1, NIST16, and DEF-12k. The main empirical claim, in Table 4, is that SparseViT achieves the best average pixel F1 (0.671) and AUC (0.937) across the first four datasets at the lowest FLOPs (46.2G) among compared methods, including TruFor and CAT-Net.

Significance. If the empirical protocol is made sound, the paper would make a useful contribution: SparseViT is architecturally simple, does not use handcrafted forensic filters, releases code, and includes ablations showing that sparse attention helps on several datasets and that the LFF head is parameter-efficient relative to MLP, ASPP, and AFF. The transfer experiments in Appendix E, showing that sparsification also improves vanilla ViT and VOLO, strengthen the generality of the mechanism. However, the central cross-dataset generalization claim is currently weakened by apparent selection of defining hyperparameters on the same test sets used for the final comparison, by the omission of DEF-12k from the main comparison table, and by the absence of any statistical significance assessment.

major comments (4)
  1. [Appendix B, Tables 3 and 6 vs Table 4] The exponential sparsity schedule in Eqs. (1)-(2) and the choice of the LFF head over single-scale or MLP fusion appear to have been selected using the same four datasets that later appear in Table 4. Appendix B reports pixel F1 for constant sparsity rates 2, 4, and 8 on COVERAGE, Columbia, CASIAv1, and NIST16, and then motivates the exponential combination from those results; Table 3 selects LFF on those same test sets. Since Table 4 reports a single run of the final model on those datasets with no held-out validation split, the reported average F1/AUC advantage may reflect test-set tuning rather than cross-dataset generalization. Please either re-run the comparison under a protocol where the schedule and fusion head are fixed before seeing the test sets, or explicitly justify why the Appendix B experiments do not constitute test-set selection; in either case, report variability across runs.
  2. [Appendix A, Tables 2 and 4] The experimental setup states that the model is tested on CASIAv1, NIST16, COVERAGE, Columbia, and DEF-12k, and Table 2 includes DEF-12k, but the state-of-the-art comparison in Table 4 omits DEF-12k entirely. This omission is consequential: Table 2 shows SparseViT's DEF-12k F1 (0.197) is lower than Uniformer with LFF alone (0.202), so the averaged claim in Table 4 is not representative of all datasets the authors claim to evaluate. Please include DEF-12k in the main comparison or state clearly why it is excluded, and adjust the generalization claim accordingly.
  3. [Introduction and Method (Sparse Self-Attention)] The paper's central mechanistic premise, that non-semantic features are 'consistent across patches unless manipulation occurs' and therefore sparse block interactions are sufficient to extract them while dense interactions are needed for semantics, is asserted rather than measured. Figure 4 is a qualitative attention visualization, and no quantitative evidence of patch independence, semantic suppression, or the local nature of non-semantic traces is provided. Since this premise motivates the entire architecture and the title's claim of 'nonsemantics-centered' extraction, please add a direct quantitative analysis (for example, feature-correlation or attention-distance statistics on pristine versus manipulated patches, or a controlled comparison of dense and sparse attention on the same features) or clearly reframe the premise as a design hypothesis that is validated only indirectly through performance.
  4. [Table 4 and Results] The main performance comparison reports single-run F1 and AUC values with no error bars or significance tests. Some margins are small (CASIAv1 F1 0.827 vs 0.818 for TruFor; NIST16 AUC 0.861 vs 0.845), and the paper claims state-of-the-art status on the basis of these averages. Please report mean and standard deviation over at least three seeds for SparseViT and, if feasible, for the compared methods, or provide a paired significance test for the main Table 4 comparisons.
minor comments (5)
  1. [Title] The title uses 'Spare-Coding Transformer' while the body consistently says 'sparse'; the intended term is presumably 'Sparse-Coding Transformer'. Also, the paper does not use sparse coding in the dictionary-learning sense, so consider whether the title should say 'Sparse-Attention Transformer' to avoid terminological confusion.
  2. [Table 2 caption] The caption says the model is 'validated on CASIAv1', but the table reports results on COVERAGE, Columbia, CASIAv1, NIST16, and DEF-12k. Please correct the caption to describe the actual evaluation sets.
  3. [Equations (1) and (2)] The notation for the sparsity-rate schedule is unclear: 'S3bi_S' and 'S4bi_S' are not defined as subscripts or superscripts, and the meaning of the superscript 'bi' is explained only loosely in the text. Please define the index notation precisely, for example S_3^{(b_i)} and S_4^{(b_i)}.
  4. [Experimental Setup / Implementation] The paper states that SparseViT is initialized with ImageNet-1k pretrained Uniformer weights. The claim 'without any handcrafted feature extractors' is accurate with respect to explicit forensic filters, but the model does inherit pretrained semantic features; please clarify this in the method or discussion so readers do not overinterpret the claim as training from scratch or as having no semantic prior.
  5. [Abstract and Table 5] The abstract states a maximum 80% FLOPs reduction, but Table 5 mixes input resolutions (256x256 for ManTraNet and PSCC-Net, 512x512 for others). A direct FLOPs comparison should use the same input resolution or report FLOPs normalized per pixel; please specify the comparison conditions for the 80% claim.

Circularity Check

1 steps flagged · score 6.0 of 10

Table 4's state-of-the-art F1 values are identical to the F1 values used in Tables 6 and 3 to choose the exponential sparsity schedule and LFF head on the same four test datasets, so the headline generalization result is a selected rather than independently predicted outcome.

  1. fitted input called prediction [Appendix B (Table 6), Table 3, Table 4]
    "we tested the model's pixel-level F1 scores under different sparsity rates (2, 4, 8) across four different datasets. ... we propose a new strategy: applying sparsification to self-attention with exponentially decreasing sparsity rates across different layers of the model. In Table 6 the row 'SparseViT 0.513 0.959 0.827 0.384' and in Table 3 the row 'LFF 0.513 0.959 0.827 0.384' are the same four-dataset F1 values later reported as 'Ours (SparseViT) 0.513 0.959 0.827 0.384' in Table 4."

    The paper's own tables show the selection step and the reported 'prediction' step are the same numbers. Appendix B (Table 6) compares constant sparsity rates 2, 4, and 8 with the full SparseViT schedule on COVERAGE, Columbia, CASIAv1, and NIST16, and then adopts the exponential schedule after seeing these test-set F1 scores. Table 3 selects LFF because it has the highest F1 on those same four datasets. Table 4 then presents the exact same four F1 values as the state-of-the-art comparison and as evidence that SparseViT is 'superior in both generalization and efficiency.' No held-out validation split separates model choice from the reported evaluation, so the headline generalization claim is the fitted selection outcome, not an independent prediction.

full rationale

The architectural derivation itself is not definitionally circular: SparseViT is trained end-to-end on the CAT-Net dataset with ground-truth masks, the sparse self-attention equations (1)-(2) do not define the target in terms of a fitted parameter, and the model is evaluated on external benchmarks. However, the headline empirical claim reduces in part to its own selection data. The exponential sparsity schedule and the LFF head are chosen using pixel-F1 on the exact four datasets that later appear in Table 4, and the final SoTA row reproduces the selection-table numbers (0.513, 0.959, 0.827, 0.384) verbatim. Thus the claimed cross-dataset generalization is not an independent prediction but a selected result; a held-out validation split or fixed hyperparameters before seeing the test sets would be needed to support the generalization claim. Self-citations to IML-ViT and IMDL-BenCo for the evaluation protocol are not load-bearing because the benchmark numbers are externally defined and the protocol is standard. The qualitative attention analysis is post hoc interpretation, not circular. Overall, one central empirical claim is circular by construction, while the proposed mechanism retains independent architectural content.

Assumptions & free parameters 3 free parameters · 4 assumptions · 0 invented entities

The central empirical claim rests on a domain assumption about the local independence of non-semantic features, a hand-selected sparsity schedule, and the standard cross-dataset evaluation protocol. No new physical or mathematical entities are introduced.

free parameters (3)
  • Sparsity-rate schedule (S in Stages 3 and 4) = exponential decay, Eq. 1 and Eq. 2 (rates from 8 down to 1)
    Hand-chosen architectural hyperparameter. Appendix B compares single rates 2, 4, 8 on the same test datasets and then adopts the mixed exponential schedule, so the schedule is effectively fit to the evaluation benchmarks.
  • LFF fusion channel width = 512
    Chosen by hand for the prediction head; minor effect on capacity.
  • LFF gamma initialization = 1e-6
    Initialized to a small value per LayerScale; learned during training, but the initial value is a hand choice.
assumptions (4)
  • domain assumption Non-semantic features are context-irrelevant, manipulation-sensitive, and consistent across image patches unless manipulation occurs.
    Stated in the Abstract and Introduction; used to justify replacing dense global attention with blockwise sparse attention.
  • domain assumption Sparse, discrete interactions among patches are sufficient to extract non-semantic features, while dense continuous interactions are required for semantics.
    Introduction and Method; the core architectural premise, not directly measured.
  • domain assumption A UniFormer backbone pre-trained on ImageNet provides a useful initialization for IML.
    Appendix A: the model is initialized with ImageNet-1k pre-trained UniFormer weights.
  • domain assumption Training on the CAT-Net joint dataset and testing on other public datasets measures generalization.
    Experimental setup in Section 'Results' and Appendix A; assumes cross-dataset transfer is the right evaluation protocol.

how reviews work

0 comments
Cite this review

Pith. "Pith review of Can We Get Rid of Handcrafted Feature Extractors? SparseViT: Nonsemantics-Centered, Parameter-Efficient Image Manipulation Localization through Spare-Coding Transformer." pith.science (2026). https://pith.science/paper/ZWYJUMOZ

@misc{pith2026241214598,
  author       = {Pith},
  title        = {Pith review of: Can We Get Rid of Handcrafted Feature Extractors? SparseViT: Nonsemantics-Centered, Parameter-Efficient Image Manipulation Localization through Spare-Coding Transformer},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/ZWYJUMOZ}},
  note         = {Machine review of arXiv:2412.14598}
}
read the original abstract

Non-semantic features or semantic-agnostic features, which are irrelevant to image context but sensitive to image manipulations, are recognized as evidential to Image Manipulation Localization (IML). Since manual labels are impossible, existing works rely on handcrafted methods to extract non-semantic features. Handcrafted non-semantic features jeopardize IML model's generalization ability in unseen or complex scenarios. Therefore, for IML, the elephant in the room is: How to adaptively extract non-semantic features? Non-semantic features are context-irrelevant and manipulation-sensitive. That is, within an image, they are consistent across patches unless manipulation occurs. Then, spare and discrete interactions among image patches are sufficient for extracting non-semantic features. However, image semantics vary drastically on different patches, requiring dense and continuous interactions among image patches for learning semantic representations. Hence, in this paper, we propose a Sparse Vision Transformer (SparseViT), which reformulates the dense, global self-attention in ViT into a sparse, discrete manner. Such sparse self-attention breaks image semantics and forces SparseViT to adaptively extract non-semantic features for images. Besides, compared with existing IML models, the sparse self-attention mechanism largely reduced the model size (max 80% in FLOPs), achieving stunning parameter efficiency and computation reduction. Extensive experiments demonstrate that, without any handcrafted feature extractors, SparseViT is superior in both generalization and efficiency across benchmark datasets.

Figures

Figures reproduced from arXiv: 2412.14598 by the authors.

Figure 1
Figure 1. SparseViT. SparseViT consists of two key components: an encoder with a sparse self-attention mechanism and a [PITH_FULL_IMAGE:figures/full_fig_p003_1.png] view at source ↗
Figure 2
Figure 2. Sparse Self-Attention. A diagram illustrating the [PITH_FULL_IMAGE:figures/full_fig_p003_2.png] view at source ↗
Figure 3
Figure 3. The Structure of LFF. By introducing learnable [PITH_FULL_IMAGE:figures/full_fig_p004_3.png] view at source ↗
Figures from the paper (5 more)
Figure 4
Figure 4. Figure 4: We select an anchor point in the manipulation re [PITH_FULL_IMAGE:figures/full_fig_p005_4.png]
Figure 5
Figure 5. Figure 5: IML by the SoTA. Existing models exhibit noticeable semantic-related false positives in the last three rows. Our model, [PITH_FULL_IMAGE:figures/full_fig_p007_5.png]
Figure 6
Figure 6. Figure 6: Robustness Analysis on CASIAv1. The results are [PITH_FULL_IMAGE:figures/full_fig_p007_6.png]
Figure 7
Figure 7. Figure 7: Qualitative Analysis Under Different Sparsity [PITH_FULL_IMAGE:figures/full_fig_p010_7.png]
Figure 8
Figure 8. Figure 8: Examples of the ability of sparse self-attention and handcrafted feature extractors to localize manipulated regions. [PITH_FULL_IMAGE:figures/full_fig_p011_8.png]

Discussion (0). Continue with ORCID to comment.

Forward citations

Cited by 1 Pith paper

Reviewed papers in the Pith corpus that reference this work. Sorted by Pith novelty score. Full citation record

  1. Unmasking Synthetic Realities in Generative AI: A Comprehensive Review of Adversarially Robust Deepfake Detection Systems

    cs.CR 2025-07 conditional novelty 3.0 of 10

    A systematic review of deepfake detection finds a pervasive lack of adversarial robustness evaluation across all modalities and calls for resilient, modality-agnostic detectors.

Reference graph

Works this paper leans on

44 extracted references · 17 canonical work pages · cited by 1 Pith paper

  1. [1]

    , " * write output.state after.block = add.period write newline

    ENTRY address archivePrefix author booktitle chapter edition editor eid eprint howpublished institution isbn journal key month note number organization pages publisher school series title type volume year label extra.label sort.label short.list INTEGERS output.state before.all mid.sentence after.sentence after.block FUNCTION init.state.consts #0 'before.a...

  2. [2]

    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 capitalize " " * FUNCT...

  3. [3]

    Arnab, A.; Dehghani, M.; Heigold, G.; Sun, C.; Lu c i \'c , M.; and Schmid, C. 2021. Vivit: A video vision transformer. In Proceedings of the IEEE/CVF international conference on computer vision, 6836--6846

    2021

  4. [4]

    Bayar, B.; and Stamm, M. C. 2018. Constrained convolutional neural networks: A new approach towards general purpose image manipulation detection. IEEE Transactions on Information Forensics and Security, 13(11): 2691--2706

    work page 2018

  5. [5]

    Chen, L.-C.; Papandreou, G.; Kokkinos, I.; Murphy, K.; and Yuille, A. L. 2017. Deeplab: Semantic image segmentation with deep convolutional nets, atrous convolution, and fully connected crfs. IEEE transactions on pattern analysis and machine intelligence, 40(4): 834--848

    2017

  6. [6]

    Child, R.; Gray, S.; Radford, A.; and Sutskever, I. 2019. Generating long sequences with sparse transformers. arXiv preprint arXiv:1904.10509

    arXiv 2019

  7. [7]

    Cozzolino, D.; and Verdoliva, L. 2019. Noiseprint: A CNN-based camera model fingerprint. IEEE Transactions on Information Forensics and Security, 15: 144--159

    work page 2019

  8. [8]

    Dai, Y.; Gieseke, F.; Oehmcke, S.; Wu, Y.; and Barnard, K. 2021. Attentional feature fusion. In Proceedings of the IEEE/CVF winter conference on applications of computer vision, 3560--3569

    work page 2021

Show all 44 references
  1. [9]

    Dong, C.; Chen, X.; Hu, R.; Cao, J.; and Li, X. 2022 a . Mvss-net: Multi-view multi-scale supervised networks for image manipulation detection. IEEE Transactions on Pattern Analysis and Machine Intelligence, 45(3): 3539--3553

    2022

  2. [10]

    Dong, J.; Wang, W.; and Tan, T. 2013. CASIA Image Tampering Detection Evaluation Database. In 2013 IEEE China Summit and International Conference on Signal and Information Processing, 422–426. Beijing, China: IEEE. ISBN 978-1-4799-1043-4

    2013

  3. [11]

    Dong, X.; Bao, J.; Chen, D.; Zhang, W.; Yu, N.; Yuan, L.; Chen, D.; and Guo, B. 2022 b . Cswin transformer: A general vision transformer backbone with cross-shaped windows. In Proceedings of the IEEE/CVF conference on computer vision and pattern recognition, 12124--12134

    2022

  4. [12]

    Dosovitskiy, A.; Beyer, L.; Kolesnikov, A.; Weissenborn, D.; Zhai, X.; Unterthiner, T.; Dehghani, M.; Minderer, M.; Heigold, G.; Gelly, S.; et al. 2020. An image is worth 16x16 words: Transformers for image recognition at scale. arXiv preprint arXiv:2010.11929

    2020 arXiv

  5. [13]

    N.; Delgado, A.; Zhou, D.; Kheyrkhah, T.; Smith, J.; and Fiscus, J

    Guan, H.; Kozak, M.; Robertson, E.; Lee, Y.; Yates, A. N.; Delgado, A.; Zhou, D.; Kheyrkhah, T.; Smith, J.; and Fiscus, J. 2019. MFC Datasets: Large-Scale Benchmark Datasets for Media Forensic Challenge Evaluation. In 2019 IEEE Winter Applications of Computer Vision Workshops ...

    2019

  6. [14]

    Guillaro, F.; Cozzolino, D.; Sud, A.; Dufour, N.; and Verdoliva, L. 2023. Trufor: Leveraging all-round clues for trustworthy image forgery detection and localization. In Proceedings of the IEEE/CVF conference on computer vision and pattern recognition, 20606--20615

    2023

  7. [15]

    Hsu, Y.-f.; and Chang, S.-f. 2006. Detecting Image Splicing using Geometry Invariants and Camera Characteristics Consistency. In 2006 IEEE International Conference on Multimedia and Expo, 549–552. Toronto, ON, Canada: IEEE. ISBN 978-1-4244-0367-7

    2006

  8. [16]

    Hu, X.; Zhang, Z.; Jiang, Z.; Chaudhuri, S.; Yang, Z.; and Nevatia, R. 2020. SPAN: Spatial pyramid attention network for image manipulation localization. In Computer Vision--ECCV 2020: 16th European Conference, Glasgow, UK, August 23--28, 2020, Proceedings, Part XXI 16, 312--3...

    2020

  9. [17]

    Kwon, M.-J.; Yu, I.-J.; Nam, S.-H.; and Lee, H.-K. 2021. CAT-Net: Compression artifact tracing network for detection and localization of image splicing. In Proceedings of the IEEE/CVF Winter Conference on Applications of Computer Vision, 375--384

    2021

  10. [18]

    Li, K.; Wang, Y.; Zhang, J.; Gao, P.; Song, G.; Liu, Y.; Li, H.; and Qiao, Y. 2023. Uniformer: Unifying convolution and self-attention for visual recognition. IEEE Transactions on Pattern Analysis and Machine Intelligence, 45(10): 12581--12600

    2023

  11. [19]

    Lin, T.-Y.; Doll \'a r, P.; Girshick, R.; He, K.; Hariharan, B.; and Belongie, S. 2017. Feature pyramid networks for object detection. In Proceedings of the IEEE conference on computer vision and pattern recognition, 2117--2125

    2017

  12. [20]

    Liu, H.; Dai, Z.; So, D.; and Le, Q. V. 2021 a . Pay attention to mlps. Advances in neural information processing systems, 34: 9204--9215

    2021

  13. [21]

    Liu, X.; Liu, Y.; Chen, J.; and Liu, X. 2022. PSCC-Net: Progressive spatio-channel correlation network for image manipulation detection and localization. IEEE Transactions on Circuits and Systems for Video Technology, 32(11): 7505--7517

    2022

  14. [22]

    Liu, Z.; Lin, Y.; Cao, Y.; Hu, H.; Wei, Y.; Zhang, Z.; Lin, S.; and Guo, B. 2021 b . Swin transformer: Hierarchical vision transformer using shifted windows. In Proceedings of the IEEE/CVF international conference on computer vision, 10012--10022

    2021

  15. [23]

    Ma, X.; Du, B.; Liu, X.; Hammadi, A. Y. A.; and Zhou, J. 2023. Iml-vit: Image manipulation localization by vision transformer. arXiv preprint arXiv:2307.14863

    2023 arXiv

  16. [24]

    Ma, X.; Zhu, X.; Su, L.; Du, B.; Jiang, Z.; Tong, B.; Lei, Z.; Yang, X.; Pun, C.-M.; Lv, J.; et al. 2024. IMDL-BenCo: A Comprehensive Benchmark and Codebase for Image Manipulation Detection & Localization. arXiv preprint arXiv:2406.10580

    2024 arXiv

  17. [25]

    L.; and Pic, M

    Mahfoudi, G.; Tajini, B.; Retraint, F.; Morain-Nicolier, F.; Dugelay, J. L.; and Pic, M. 2019. DEFACTO: Image and Face Manipulation Dataset. In 2019 27th European Signal Processing Conference (EUSIPCO), 1–5. A Coruna, Spain: IEEE. ISBN 978-90-827970-3-9

    2019

  18. [26]

    Pun, C.-M.; Yuan, X.-C.; and Bi, X.-L. 2015. Image forgery detection using adaptive oversegmentation and feature point matching. ieee transactions on information forensics and security, 10(8): 1705--1716

    2015

  19. [27]

    Simonyan, K.; and Zisserman, A. 2014. Very deep convolutional networks for large-scale image recognition. arXiv preprint arXiv:1409.1556

    2014 arXiv

  20. [28]

    Touvron, H.; Bojanowski, P.; Caron, M.; Cord, M.; El-Nouby, A.; Grave, E.; Izacard, G.; Joulin, A.; Synnaeve, G.; Verbeek, J.; et al. 2022. Resmlp: Feedforward networks for image classification with data-efficient training. IEEE transactions on pattern analysis and machine int...

    2022

  21. [29]

    Touvron, H.; Cord, M.; Sablayrolles, A.; Synnaeve, G.; and J \'e gou, H. 2021. Going deeper with image transformers. In Proceedings of the IEEE/CVF international conference on computer vision, 32--42

    2021

  22. [30]

    Ulyanov, D.; Vedaldi, A.; and Lempitsky, V. 2018. Deep image prior. In Proceedings of the IEEE conference on computer vision and pattern recognition, 9446--9454

    2018

  23. [31]

    Vaswani, A. 2017. Attention is all you need. arXiv preprint arXiv:1706.03762

    2017 arXiv

  24. [32]

    Wang, J.; Sun, K.; Cheng, T.; Jiang, B.; Deng, C.; Zhao, Y.; Liu, D.; Mu, Y.; Tan, M.; Wang, X.; et al. 2020. Deep high-resolution representation learning for visual recognition. IEEE transactions on pattern analysis and machine intelligence, 43(10): 3349--3364

    2020

  25. [33]

    Wang, J.; Wu, Z.; Chen, J.; Han, X.; Shrivastava, A.; Lim, S.-N.; and Jiang, Y.-G. 2022. Objectformer for image manipulation detection and localization. In Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition, 2364--2373

    2022

  26. [34]

    Wang, W.; Xie, E.; Li, X.; Fan, D.-P.; Song, K.; Liang, D.; Lu, T.; Luo, P.; and Shao, L. 2021. Pyramid vision transformer: A versatile backbone for dense prediction without convolutions. In Proceedings of the IEEE/CVF international conference on computer vision, 568--578

    2021

  27. [35]

    Wang, X.; Girshick, R.; Gupta, A.; and He, K. 2018. Non-local neural networks. In Proceedings of the IEEE conference on computer vision and pattern recognition, 7794--7803

    2018

  28. [36]

    Wei, Y.; Xiao, B.; Bi, X.; Ma, Z.; Liu, Y.; and Ma, Z. 2023. Secondary Labeling: A Novel Labeling Strategy for Image Manipulation Detection. In Proceedings of the 31st ACM International Conference on Multimedia, 8225--8232

    2023

  29. [37]

    Wen, B.; Zhu, Y.; Subramanian, R.; Ng, T.-T.; Shen, X.; and Winkler, S. 2016. COVERAGE — A novel database for copy-move forgery detection. In 2016 IEEE International Conference on Image Processing (ICIP), 161–165. Phoenix, AZ, USA: IEEE. ISBN 978-1-4673-9961-6

    2016

  30. [38]

    Wu, Y.; AbdAlmageed, W.; and Natarajan, P. 2019. Mantra-net: Manipulation tracing network for detection and localization of image forgeries with anomalous features. In Proceedings of the IEEE/CVF conference on computer vision and pattern recognition, 9543--9552

    2019

  31. [39]

    M.; and Luo, P

    Xie, E.; Wang, W.; Yu, Z.; Anandkumar, A.; Alvarez, J. M.; and Luo, P. 2021. SegFormer: Simple and efficient design for semantic segmentation with transformers. Advances in neural information processing systems, 34: 12077--12090

    2021

  32. [40]

    Yang, J.; Li, C.; Zhang, P.; Dai, X.; Xiao, B.; Yuan, L.; and Gao, J. 2021. Focal self-attention for local-global interactions in vision transformers. arXiv preprint arXiv:2107.00641

    2021 arXiv

  33. [41]

    E.; Feng, J.; and Yan, S

    Yuan, L.; Chen, Y.; Wang, T.; Yu, W.; Shi, Y.; Jiang, Z.-H.; Tay, F. E.; Feng, J.; and Yan, S. 2021. Tokens-to-token vit: Training vision transformers from scratch on imagenet. In Proceedings of the IEEE/CVF international conference on computer vision, 558--567

    2021

  34. [42]

    Yuan, L.; Hou, Q.; Jiang, Z.; Feng, J.; and Yan, S. 2022. Volo: Vision outlooker for visual recognition. IEEE transactions on pattern analysis and machine intelligence, 45(5): 6575--6586

    2022

  35. [43]

    Y.; and Feng, W

    Zhou, J.; Ma, X.; Du, X.; Alhammadi, A. Y.; and Feng, W. 2023. Pre-training-free image manipulation localization through non-mutually exclusive contrastive learning. In Proceedings of the IEEE/CVF International Conference on Computer Vision, 22346--22356

    2023

  36. [44]

    I.; and Davis, L

    Zhou, P.; Han, X.; Morariu, V. I.; and Davis, L. S. 2018. Learning rich features for image manipulation detection. In Proceedings of the IEEE conference on computer vision and pattern recognition, 1053--1061

    2018

Pith tools

Reviewed August 11, 2026 · model on record in the stance chip above.