pith. machine review for the scientific record. sign in

arxiv: 2604.23016 · v1 · submitted 2026-04-24 · 💻 cs.CR · cs.AI· cs.CV

Recognition: unknown

DeepSignature: Digitally Signed, Content-Encoding Watermarks for Robust and Transparent Image Authentication

Mathias Graf , Marco Willi , Melanie Mathys , Michael Aerni , Christian Schwarzer , Martin Melchior , Michael H. Graber
Authors on Pith no claims yet

Pith reviewed 2026-05-08 11:25 UTC · model grok-4.3

classification 💻 cs.CR cs.AIcs.CV
keywords digital watermarkingimage authenticationneural networksdigital signaturestamper detectionforgery detectioncontent integritylatent space verification
0
0 comments X

The pith

Neural networks can embed cryptographically signed content-encoding watermarks into images to verify origin and detect tampering without special formats.

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

The paper presents DeepSignature as a way to use neural networks to hide content-encoding watermarks in images that also carry digital signatures. These allow verification of where the image came from and whether it has been tampered with, all without changing how images are stored or viewed. This matters because easy image editing with AI threatens to undermine trust in photos from news, social media, or official sources. The system adds a way to check for changes inside the neural network's feature space. Tests indicate it can catch most serious forgeries while handling everyday image adjustments.

Core claim

DeepSignature integrates digital signatures with deep neural networks by using the networks to generate content-encoding watermarks and embed them imperceptibly into images. The watermarks are cryptographically verifiable, enabling source attribution and image integrity validation. The system is compatible with existing image formats, supports client-side verification with the signer's public key, and includes a novel latent-space verification approach to detect and localize tampering. Experiments confirm near 100% identification of significant forgery attempts.

What carries the argument

Neural networks that generate and embed imperceptible content-encoding watermarks combined with cryptographic digital signatures for verification and latent-space analysis for tamper localization.

If this is right

  • Standard image files can carry proof of authenticity and origin without any format changes or extra files.
  • Tampering attempts can be detected and localized using internal network features even when the image looks normal.
  • Parameters can be tuned to balance imperceptibility against robustness for specific use cases.
  • Verification requires only the signer's public key and runs on the client side.
  • Common edits like compression are tolerated while substantial forgeries are flagged reliably.

Where Pith is reading between the lines

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

  • The same embedding idea could apply to video frames or audio to authenticate other media types.
  • Pairing the public keys with a distributed ledger might enable verification without relying on any single trusted party.
  • The latent-space localization could be extended to classify the type of edit performed, such as object insertion or removal.
  • Attackers might try to train against the specific embedding network to produce images that bypass detection.

Load-bearing premise

Neural networks can simultaneously create watermarks that stay invisible to the eye, survive transformations and attacks, carry verifiable cryptographic signatures, and support accurate tamper localization in latent space across diverse images.

What would settle it

A forged image whose significant content changes still extract to a valid matching signature, or a minor non-content change that causes a legitimate signed image to fail verification.

Figures

Figures reproduced from arXiv: 2604.23016 by Christian Schwarzer, Marco Willi, Martin Melchior, Mathias Graf, Melanie Mathys, Michael Aerni, Michael H. Graber.

Figure 1
Figure 1. Figure 1: DeepSignature reliably detects and localizes tampering in signed images. The figure presents results for view at source ↗
Figure 2
Figure 2. Figure 2: Illustration of DeepSignature. The left panel illustrates the signer’s workflow, which involves encoding an view at source ↗
Figure 3
Figure 3. Figure 3: DeepSignature introduces more uniformly distributed noise artifacts compared to JPEG compression. The view at source ↗
Figure 4
Figure 4. Figure 4: DeepSignature remains robust to various types of benign image transformations. The figure shows panels for view at source ↗
Figure 5
Figure 5. Figure 5: Under JPEG compression (quality 50), DeepSignature supports substantially higher payload capacity than the view at source ↗
Figure 6
Figure 6. Figure 6: Latent-space verification accurately highlights tampered regions without being influenced by spurious view at source ↗
Figure 8
Figure 8. Figure 8: Payload composition for an example image of size 1 view at source ↗
Figure 9
Figure 9. Figure 9: Some Emu Edit forgeries are barely perceptible. Shown are example image pairs from Emu Edit that were view at source ↗
read the original abstract

AI-powered generative models have significantly expanded the possibilities for editing, manipulating, and creating high-quality images. Particularly, images that falsely appear to originate from trusted sources pose a serious threat, undermining public trust in image authenticity. We propose DeepSignature, a novel approach that integrates the guarantees of digital signatures with the capabilities of deep neural networks. Neural networks are used both to generate content-encoding watermarks and to embed them imperceptibly into images while ensuring robust extraction. These watermarks are cryptographically verifiable, enabling source attribution and image integrity validation. DeepSignature is compatible with existing image formats and requires no special handling of signed images. It supports client-side verification, requiring only the signer's public key. Additionally, we introduce a novel latent-space verification approach to detect and localize tampering attempts. We evaluate DeepSignature in terms of imperceptibility, robustness to benign transformations, forgery detection, and its resilience against various attack scenarios. Our results highlight the inherent trade-offs between imperceptibility, robustness, and integrity verification. We demonstrate that DeepSignature reliably identifies significant forgery attempts -- achieving near 100\% in our experiments. Finally, we emphasize DeepSignature's modularity and tunable parameters, allowing adaptation to application-specific requirements. Code and model weights will be published.

Editorial analysis

A structured set of objections, weighed in public.

Desk editor's note, referee report, simulated authors' rebuttal, and a circularity audit. Tearing a paper down is the easy half of reading it; the pith above is the substance, this is the friction.

Referee Report

2 major / 0 minor

Summary. The manuscript proposes DeepSignature, a system that uses neural networks to generate content-encoding watermarks and embed them imperceptibly into images, combined with cryptographic digital signatures for source attribution and integrity verification. It supports standard image formats, client-side verification via public key, and introduces a latent-space approach for tamper detection and localization. The work claims robustness to benign transformations (JPEG, resize, etc.), near-100% detection of significant forgeries, and tunable parameters for imperceptibility-robustness trade-offs, with plans to release code and weights.

Significance. If the performance claims are substantiated with detailed metrics, this hybrid crypto-ML approach could provide a deployable method for authenticating images against AI-generated manipulations, with advantages in format compatibility and localization. The modularity and open release plans are strengths. However, without quantitative results or baselines, its significance relative to prior watermarking and provenance schemes cannot yet be determined.

major comments (2)
  1. Abstract: The central claim that DeepSignature 'reliably identifies significant forgery attempts -- achieving near 100% in our experiments' is stated without any quantitative metrics, tables, figures, baselines, attack details, error bars, or descriptions of the evaluation protocol. This prevents assessment of the forgery detection performance and the overall soundness of the results.
  2. Verification mechanism (described in abstract and architecture): Digital signatures require bit-exact recovery of the signed content-encoding watermark. The neural network extractor must therefore achieve zero bit-error rate on authentic images after allowed transformations, yet the manuscript provides no details on error-correcting codes, repetition, quantization, or other mechanisms to enforce exactness while preserving imperceptibility and enabling the latent-space localization path. Without this, verification would fail on any extraction error, rendering the forgery-detection metric uninterpretable.

Simulated Author's Rebuttal

2 responses · 0 unresolved

We thank the referee for the constructive feedback and the recommendation for major revision. We address each major comment point by point below, committing to revisions that strengthen the manuscript without misrepresenting our current results.

read point-by-point responses

  1. Referee: Abstract: The central claim that DeepSignature 'reliably identifies significant forgery attempts -- achieving near 100% in our experiments' is stated without any quantitative metrics, tables, figures, baselines, attack details, error bars, or descriptions of the evaluation protocol. This prevents assessment of the forgery detection performance and the overall soundness of the results.

    Authors: We agree that the abstract claim requires quantitative substantiation to permit proper evaluation. The current manuscript's evaluation section reports forgery detection results, but these details are not summarized in the abstract. In the revised version we will update the abstract to include key metrics (e.g., exact detection rates under specified tampering scenarios), reference the evaluation protocol, baselines, and attack details, and note the presence of error bars. This change will make the performance claims directly assessable. revision: yes

  2. Referee: Verification mechanism (described in abstract and architecture): Digital signatures require bit-exact recovery of the signed content-encoding watermark. The neural network extractor must therefore achieve zero bit-error rate on authentic images after allowed transformations, yet the manuscript provides no details on error-correcting codes, repetition, quantization, or other mechanisms to enforce exactness while preserving imperceptibility and enabling the latent-space localization path. Without this, verification would fail on any extraction error, rendering the forgery-detection metric uninterpretable.

    Authors: We acknowledge that the manuscript does not currently provide explicit details on achieving bit-exact recovery. We will revise the architecture and methodology sections to describe the error-correcting codes, quantization steps, and any repetition mechanisms employed in the extractor. These additions will explain how zero bit-error rate is maintained for authentic images under allowed transformations while keeping the latent-space path available for tamper localization, thereby clarifying the interpretability of the forgery-detection results. revision: yes

Circularity Check

0 steps flagged

No significant circularity; empirical system proposal with independent experimental validation

full rationale

The paper describes an engineering system that combines standard digital signatures with trained neural networks for imperceptible watermark embedding and extraction. No derivation chain, equations, or first-principles results are presented that reduce to fitted parameters by construction, self-citations, or renamed known results. Claims of near-100% forgery detection are explicitly tied to experimental outcomes on specific datasets and attacks, not to any tautological prediction. The architecture relies on external cryptographic primitives and standard NN training, with no load-bearing step that is self-definitional or statistically forced. This matches the default expectation of a non-circular paper.

Axiom & Free-Parameter Ledger

1 free parameters · 1 axioms · 0 invented entities

The approach depends on the existence of secure digital signature schemes and on the ability of neural networks to learn imperceptible yet extractable embeddings; no new physical or mathematical entities are postulated.

free parameters (1)
  • tunable parameters for imperceptibility-robustness trade-off
    The abstract states that the system is modular with tunable parameters that can be adapted to application-specific requirements.
axioms (1)
  • standard math Standard cryptographic assumptions of digital signature unforgeability and verifiability
    The security guarantees for source attribution and integrity rest on these well-established properties of public-key signatures.

pith-pipeline@v0.9.0 · 5544 in / 1298 out tokens · 41788 ms · 2026-05-08T11:25:45.639380+00:00 · methodology

discussion (0)

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

Reference graph

Works this paper leans on

43 extracted references · 30 canonical work pages · 2 internal anchors

  1. [1]

    In: IEEE/CVF Conference on Computer Vision and Pattern Recognition, CVPR 2022, New Orleans, LA, USA, June 18-24, 2022

    Robin Rombach, Andreas Blattmann, Dominik Lorenz, Patrick Esser, and Bjorn Ommer. High-Resolution Image Synthesis with Latent Diffusion Models. In2022 IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR), pages 10674–10685, New Orleans, LA, USA, June 2022. IEEE. ISBN 978-1-6654-6946-3. doi:10.1109/CVPR52688.2022.01042. URLhttps://ieeexplo...

    work page doi:10.1109/cvpr52688.2022.01042 2022

  2. [2]

    URL https://c2pa.org/specifications/ specifications/1.0/specs/C2PA_Specification.html

    C2PA Technical Specification, December 2021. URL https://c2pa.org/specifications/ specifications/1.0/specs/C2PA_Specification.html

    2021

  3. [3]

    Friedman

    G.L. Friedman. The trustworthy digital camera: restoring credibility to the photographic image.IEEE Transactions on Consumer Electronics, 39(4):905–910, November 1993. ISSN 00983063. doi:10.1109/30.267415. URL http://ieeexplore.ieee.org/document/267415/

    work page doi:10.1109/30.267415 1993

  4. [4]

    Digital image integrity – a survey of protection and verification techniques.Digital Signal Processing, 71:1–26, December 2017

    Paweł Korus. Digital image integrity – a survey of protection and verification techniques.Digital Signal Processing, 71:1–26, December 2017. ISSN 10512004. doi:10.1016/j.dsp.2017.08.009. URL https:// linkinghub.elsevier.com/retrieve/pii/S1051200417301938

    work page doi:10.1016/j.dsp.2017.08.009 2017

  5. [5]

    Determining Image Origin and Integrity Using Sensor Noise.IEEE Transactions on Information Forensics and Security, 3(1):74–90, 2008

    Mo Chen, Jessica Fridrich, Miroslav Goljan, and Jan Lukas. Determining Image Origin and Integrity Using Sensor Noise.IEEE Transactions on Information Forensics and Security, 3(1):74–90, 2008. ISSN 1556-6013. doi:10.1109/TIFS.2007.916285. URLhttp://ieeexplore.ieee.org/document/4451084/

    work page doi:10.1109/tifs.2007.916285 2008

  6. [6]

    In: 2023 IEEE/CVF Conference on Com- puter Vision and Pattern Recognition (CVPR), pp

    Fabrizio Guillaro, Davide Cozzolino, Avneesh Sud, Nicholas Dufour, and Luisa Verdoliva. TruFor: Leveraging All-Round Clues for Trustworthy Image Forgery Detection and Localization. In2023 IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR), pages 20606–20615, Vancouver, BC, Canada, June 2023. IEEE. ISBN 979-8-3503-0129-8. doi:10.1109/CVP...

    work page doi:10.1109/cvpr52729.2023.01974 2023

  7. [7]

    Sheng-Yu Wang, Oliver Wang, Richard Zhang, Andrew Owens, and Alexei A. Efros. CNN-generated images are surprisingly easy to spot... for now, April 2020. URLhttp://arxiv.org/abs/1912.11035. arXiv:1912.11035 [cs]

    work page arXiv 2020

  8. [8]

    In: IEEE/CVF International Conference on Computer Vision, ICCV 2023 - Workshops, Paris, France, October 2-6, 2023

    David C. Epstein, Ishan Jain, Oliver Wang, and Richard Zhang. Online Detection of AI-Generated Images. In2023 IEEE/CVF International Conference on Computer Vision Workshops (ICCVW), pages 382–392, Paris, France, October 2023. IEEE. doi:10.1109/ICCVW60793.2023.00045. URL https://ieeexplore.ieee.org/ document/10350523/

    work page doi:10.1109/iccvw60793.2023.00045 2023

  9. [9]

    Ancuti, C

    Nicholas Carlini and Hany Farid. Evading Deepfake-Image Detectors with White- and Black-Box Attacks. In 2020 IEEE/CVF Conference on Computer Vision and Pattern Recognition Workshops (CVPRW), pages 2804–2813, Seattle, W A, USA, June 2020. IEEE. ISBN 978-1-7281-9360-1. doi:10.1109/CVPRW50498.2020.00337. URL https://ieeexplore.ieee.org/document/9150604/

    work page doi:10.1109/cvprw50498.2020.00337 2020

  10. [10]

    Cox, Matthew L

    Ingemar J. Cox, Matthew L. Miller, Jeffrey A. Bloom, Jessica Fridrich, and Ton Kalker.Digital Watermarking and Steganography, volume The Morgan Kaufmann Series in Multimedia Information and Systems. Elsevier, 2nd edition, 2008. ISBN 978-0-12-372585-1. doi:10.1016/B978-0-12-372585-1.X5001-3. URL https:// linkinghub.elsevier.com/retrieve/pii/B9780123725851X50013

    work page doi:10.1016/b978-0-12-372585-1.x5001-3 2008

  11. [11]

    Chi-Kwong Chan and L.M. Cheng. Hiding data in images by simple LSB substitution.Pattern Recognition, 37 (3):469–474, March 2004. ISSN 00313203. doi:10.1016/j.patcog.2003.08.007. URL https://linkinghub. elsevier.com/retrieve/pii/S003132030300284X

    work page doi:10.1016/j.patcog.2003.08.007 2004

  12. [12]

    A Review on Detection of LSB Matching Steganography.Infor- mation Technology Journal, 9(8):1725–1738, November 2010

    Jiaohua Qin, Xuyu Xiang, and Meng Xian Wang. A Review on Detection of LSB Matching Steganography.Infor- mation Technology Journal, 9(8):1725–1738, November 2010. ISSN 18125638. doi:10.3923/itj.2010.1725.1738. URLhttps://www.scialert.net/abstract/?doi=itj.2010.1725.1738

    work page doi:10.3923/itj.2010.1725.1738 2010

  13. [13]

    Confidence in Assurance 2.0 Cases

    Jiren Zhu, Russell Kaplan, Justin Johnson, and Li Fei-Fei. HiDDeN: Hiding Data With Deep Networks. InCom- puter Vision – ECCV 2018: 15th European Conference, Munich, Germany, September 8-14, 2018, Proceedings, Part XV, pages 682–697, Berlin, Heidelberg, 2018. Springer-Verlag. ISBN 978-3-030-01266-3. doi:10.1007/978- 3-030-01267-0_40. URLhttps://doi.org/10...

    work page doi:10.1007/978- 2018

  14. [14]

    Hiding Images in Plain Sight: Deep Steganography

    Shumeet Baluja. Hiding Images in Plain Sight: Deep Steganography. InProceedings of the 31st International Conference on Neural Information Processing Systems, NIPS’17, pages 2066–2076, Red Hook, NY , USA, 2017. Curran Associates Inc. ISBN 978-1-5108-6096-4

    2066

  15. [15]

    Cornia, M

    Xiyang Luo, Ruohan Zhan, Huiwen Chang, Feng Yang, and Peyman Milanfar. Distortion Agnostic Deep Watermarking. In2020 IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR), pages 13545– 13554, Seattle, W A, USA, June 2020. IEEE. ISBN 978-1-7281-7168-5. doi:10.1109/CVPR42600.2020.01356. URLhttps://ieeexplore.ieee.org/document/9157561/. 18 De...

    work page doi:10.1109/cvpr42600.2020.01356 2020

  16. [16]

    Cornia, M

    Matthew Tancik, Ben Mildenhall, and Ren Ng. StegaStamp: Invisible Hyperlinks in Physical Photographs. In 2020 IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR), pages 2114–2123, Seattle, WA, USA, June 2020. IEEE. ISBN 978-1-7281-7168-5. doi:10.1109/CVPR42600.2020.00219. URL https: //ieeexplore.ieee.org/document/9156548/

    work page doi:10.1109/cvpr42600.2020.00219 2020

  17. [17]

    TrustMark: Universal Watermarking for Arbitrary Resolution Images, November 2023

    Tu Bui, Shruti Agarwal, and John Collomosse. TrustMark: Universal Watermarking for Arbitrary Resolution Images, November 2023. URLhttp://arxiv.org/abs/2311.18297. arXiv:2311.18297 [cs]

    work page arXiv 2023

  18. [18]

    MBRS: Enhancing Robustness of DNN-based Watermarking by Mini-Batch of Real and Simulated JPEG Compression

    Zhaoyang Jia, Han Fang, and Weiming Zhang. MBRS: Enhancing Robustness of DNN-based Watermarking by Mini-Batch of Real and Simulated JPEG Compression. InProceedings of the 29th ACM International Conference on Multimedia, MM ’21, pages 41–49, New York, NY , USA, October 2021. Association for Computing Machinery. ISBN 978-1-4503-8651-7. doi:10.1145/3474085.3...

    work page doi:10.1145/3474085.3475324 2021

  19. [19]

    InvisMark: Invisible and Robust Watermarking for AI-generated Image Provenance, November 2024

    Rui Xu, Mengya Hu, Deren Lei, Yaxi Li, David Lowe, Alex Gorevski, Mingyu Wang, Emily Ching, and Alex Deng. InvisMark: Invisible and Robust Watermarking for AI-generated Image Provenance, November 2024. URL http://arxiv.org/abs/2411.07795. arXiv:2411.07795 [cs]

    work page arXiv 2024

  20. [20]

    Chen, Stefano Mangini, and Marcel Worring

    Pierre Fernandez, Alexandre Sablayrolles, Teddy Furon, Herve Jegou, and Matthijs Douze. Watermarking Images in Self-Supervised Latent Spaces. InICASSP 2022 - 2022 IEEE International Conference on Acoustics, Speech and Signal Processing (ICASSP), pages 3054–3058, Singapore, Singapore, May 2022. IEEE. ISBN 978-1-6654-0540-9. doi:10.1109/ICASSP43922.2022.974...

    work page doi:10.1109/icassp43922.2022.9746058 2022

  21. [21]

    Within-Camera Multilayer Perceptron DVS Denoising,

    Tu Bui, Shruti Agarwal, Ning Yu, and John Collomosse. RoSteALS: Robust Steganography using Autoen- coder Latent Space. In2023 IEEE/CVF Conference on Computer Vision and Pattern Recognition Work- shops (CVPRW), pages 933–942, Vancouver, BC, Canada, June 2023. IEEE. ISBN 979-8-3503-0249-3. doi:10.1109/CVPRW59228.2023.00100. URLhttps://ieeexplore.ieee.org/do...

    work page doi:10.1109/cvprw59228.2023.00100 2023

  22. [22]

    SWIFT: Semantic Watermarking for Image Forgery Thwarting

    Gautier Evennou, Vivien Chappelier, Ewa Kijak, and Teddy Furon. SWIFT: Semantic Watermarking for Image Forgery Thwarting. In2024 IEEE International Workshop on Information Forensics and Security (WIFS), pages 1–6, Rome, Italy, December 2024. IEEE. ISBN 979-8-3503-6442-2. doi:10.1109/WIFS61860.2024.10810692. URLhttps://ieeexplore.ieee.org/document/10810692/

    work page doi:10.1109/wifs61860.2024.10810692 2024

  23. [23]

    High-fidelity generative image compression

    Fabian Mentzer, George Toderici, Michael Tschannen, and Eirikur Agustsson. High-fidelity generative image compression. InProceedings of the 34th International Conference on Neural Information Processing Systems, NIPS ’20, Red Hook, NY , USA, 2020. Curran Associates Inc. ISBN 978-1-7138-2954-6

    2020

  24. [24]

    Learning transferable visual models from natural language supervision

    Alec Radford, Jong Wook Kim, Chris Hallacy, Aditya Ramesh, Gabriel Goh, Sandhini Agarwal, Girish Sastry, Amanda Askell, Pamela Mishkin, Jack Clark, and others. Learning transferable visual models from natural language supervision. InInternational conference on machine learning, pages 8748–8763. PMLR, 2021

    2021

  25. [25]

    doi: 10.1145/3531146.3533079

    Lukas Struppek, Dominik Hintersdorf, Daniel Neider, and Kristian Kersting. Learning to Break Deep Perceptual Hashing: The Use Case NeuralHash. In2022 ACM Conference on Fairness, Accountability, and Transparency, pages 58–69, Seoul Republic of Korea, June 2022. ACM. ISBN 978-1-4503-9352-2. doi:10.1145/3531146.3533073. URLhttps://dl.acm.org/doi/10.1145/3531...

    work page doi:10.1145/3531146.3533073 2022

  26. [26]

    Exploring the Limits of Weakly Supervised Pretraining

    Dhruv Mahajan, Ross Girshick, Vignesh Ramanathan, Kaiming He, Manohar Paluri, Yixuan Li, Ashwin Bharambe, and Laurens Van Der Maaten. Exploring the Limits of Weakly Supervised Pretraining. In Vittorio Ferrari, Martial Hebert, Cristian Sminchisescu, and Yair Weiss, editors,Computer Vision – ECCV 2018, volume 11206, pages 185–201. Springer International Pub...

    work page doi:10.1007/978-3-030-01216-8_12 2018

  27. [27]

    Series Title: Lecture Notes in Computer Science

  28. [28]

    Auto-Encoding Variational Bayes

    Diederik P. Kingma and Max Welling. Auto-Encoding Variational Bayes. In Yoshua Bengio and Yann LeCun, editors,2nd International Conference on Learning Representations, ICLR 2014, Banff, AB, Canada, April 14-16, 2014, Conference Track Proceedings, 2014. URLhttp://arxiv.org/abs/1312.6114

    work page internal anchor Pith review arXiv 2014

  29. [29]

    A simple framework for contrastive learning of visual representations

    Ting Chen, Simon Kornblith, Mohammad Norouzi, and Geoffrey Hinton. A simple framework for contrastive learning of visual representations. InProceedings of the 37th International Conference on Machine Learning, ICML’20. JMLR.org, 2020

    2020

  30. [30]

    Neural discrete representation learning

    Aaron van den Oord, Oriol Vinyals, and Koray Kavukcuoglu. Neural discrete representation learning. In Proceedings of the 31st International Conference on Neural Information Processing Systems, NIPS’17, pages 6309–6318, Red Hook, NY , USA, December 2017. Curran Associates Inc. ISBN 978-1-5108-6096-4

    2017

  31. [31]

    Edwards-Curve Digital Signature Algorithm (EdDSA)

    Simon Josefsson and Ilari Liusvaara. Edwards-Curve Digital Signature Algorithm (EdDSA). Request for Comments RFC 8032, Internet Engineering Task Force, January 2017. URL https://datatracker.ietf. org/doc/rfc8032/. Num Pages: 60. 19 DeepSignatureA PREPRINT

    2017

  32. [32]

    Recommendation for Key Management Part 1: General

    Elaine Barker. Recommendation for Key Management Part 1: General. Technical Report NIST SP 800-57pt1r4, National Institute of Standards and Technology, January 2016. URL https://nvlpubs.nist.gov/nistpubs/ SpecialPublications/NIST.SP.800-57pt1r4.pdf

    2016

  33. [33]

    R. C. Bose and D. K. Ray-Chaudhuri. On a class of error correcting binary group codes.Information and Control, 3(1):68–79, March 1960. ISSN 0019-9958. doi:10.1016/S0019-9958(60)90287-4. URL https: //www.sciencedirect.com/science/article/pii/S0019995860902874

    work page doi:10.1016/s0019-9958(60)90287-4 1960

  34. [34]

    IEEE Trans

    Z. Wang, A.C. Bovik, H.R. Sheikh, and E.P. Simoncelli. Image Quality Assessment: From Error Visibility to Structural Similarity.IEEE Transactions on Image Processing, 13(4):600–612, April 2004. ISSN 1057-7149. doi:10.1109/TIP.2003.819861. URLhttp://ieeexplore.ieee.org/document/1284395/

    work page doi:10.1109/tip.2003.819861 2004

  35. [35]

    Lawrence Zitnick, and Piotr Dollár

    Tsung-Yi Lin, Michael Maire, Serge Belongie, Lubomir Bourdev, Ross Girshick, James Hays, Pietro Perona, Deva Ramanan, C. Lawrence Zitnick, and Piotr Dollár. Microsoft COCO: Common Objects in Context, February

  36. [36]

    Microsoft COCO: Common Objects in Context

    URLhttp://arxiv.org/abs/1405.0312. arXiv:1405.0312 [cs]

    work page internal anchor Pith review arXiv

  37. [37]

    Wang, E.P

    Z. Wang, E.P. Simoncelli, and A.C. Bovik. Multiscale structural similarity for image quality assessment. In The Thrity-Seventh Asilomar Conference on Signals, Systems & Computers, 2003, pages 1398–1402, Pacific Grove, CA, USA, 2003. IEEE. ISBN 978-0-7803-8104-9. doi:10.1109/ACSSC.2003.1292216. URL http: //ieeexplore.ieee.org/document/1292216/

    work page doi:10.1109/acssc.2003.1292216 2003

  38. [38]

    In: 2018 IEEE Conference on Computer Vision and Pattern Recognition, CVPR 2018, Salt Lake City, UT, USA, June 18-22, 2018

    Richard Zhang, Phillip Isola, Alexei A. Efros, Eli Shechtman, and Oliver Wang. The Unreasonable Effectiveness of Deep Features as a Perceptual Metric. In2018 IEEE/CVF Conference on Computer Vision and Pattern Recognition, pages 586–595, Salt Lake City, UT, June 2018. IEEE. ISBN 978-1-5386-6420-9. doi:10.1109/CVPR.2018.00068. URLhttps://ieeexplore.ieee.org...

    work page doi:10.1109/cvpr.2018.00068 2018

  39. [39]

    CLIC 2024 - 6th Challenge on Learned Image Compression, January 2023

    Johannes Ballé, George Toderici, Luca Versari, Nick Johnston, Lucas Theis, Andrey Norkin, Krishna Rapaka, Erfan Noury, Ross Cutler, Radu Timofte, Fabien Racapé, Yan Ye, Ali Bilgin, Andrew Segall, Balu Adsumilli, and Ramzi Khsib. CLIC 2024 - 6th Challenge on Learned Image Compression, January 2023. URL http: //www.compression.cc/

    2024

  40. [40]

    CASIA Image Tampering Detection Evaluation Database

    Jing Dong, Wei Wang, and Tieniu Tan. CASIA Image Tampering Detection Evaluation Database. In2013 IEEE China Summit and International Conference on Signal and Information Processing, pages 422–426, Beijing, China, July 2013. IEEE. ISBN 978-1-4799-1043-4. doi:10.1109/ChinaSIP.2013.6625374. URL http://ieeexplore.ieee.org/document/6625374/

    work page doi:10.1109/chinasip.2013.6625374 2013

  41. [41]

    Vbench: Comprehensive benchmark suite for video generative models

    Shelly Sheynin, Adam Polyak, Uriel Singer, Yuval Kirstain, Amit Zohar, Oron Ashual, Devi Parikh, and Yaniv Taigman. Emu Edit: Precise Image Editing via Recognition and Generation Tasks. In2024 IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR), pages 8871–8879, Seattle, WA, USA, June 2024. IEEE. ISBN 979-8-3503-5300-6. doi:10.1109/CVPR5...

    work page doi:10.1109/cvpr52733.2024.00847 2024

  42. [42]

    DiffForensics: Leveraging Diffusion Prior to Image Forgery Detection and Localization

    Zeqin Yu, Jiangqun Ni, Yuzhen Lin, Haoyi Deng, and Bin Li. DiffForensics: Leveraging Diffusion Prior to Image Forgery Detection and Localization. InProceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR), pages 12765–12774, June 2024

    2024

  43. [43]

    UnionFormer: Unified- Learning Transformer with Multi-View Representation for Image Manipulation Detection and Localization

    Shuaibo Li, Wei Ma, Jianwei Guo, Shibiao Xu, Benchong Li, and Xiaopeng Zhang. UnionFormer: Unified- Learning Transformer with Multi-View Representation for Image Manipulation Detection and Localization. InProceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR), pages 12523–12533, June 2024. 20

    2024