pith. machine review for the scientific record. sign in

arxiv: 2605.04705 · v1 · submitted 2026-05-06 · 💻 cs.CR · cs.LG

Recognition: unknown

Vol-Mark: A Watermark for 3D Medical Volume Data Via Cubic Difference Expansion and Contrastive Learning

Jiangnan Zhu , Yuntao Wang , Shengli Pan , Yujie Gu
Authors on Pith no claims yet

Pith reviewed 2026-05-08 17:56 UTC · model grok-4.3

classification 💻 cs.CR cs.LG
keywords 3D medical volumereversible watermarkingcontrastive learningdifference expansionwavelet transformownership verificationtelemedicine securitydata integrity
0
0 comments X

The pith

Vol-Mark watermarks 3D medical volume data using contrastive features and cubic difference expansions to enable reversible ownership protection with high attack resistance.

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

The paper presents Vol-Mark as a reversible-zero watermarking scheme tailored for 3D medical volumes that are shared in telemedicine. It first trains a contrastive learning model to pull out stable volumetric features that can identify the data even after it has been altered. It then applies a 3D integer wavelet transform and expands voxel differences inside small cubes at the low-frequency coefficients to hide the watermark bits with minimal change to the original data. A majority vote during extraction improves reliability, and the whole mark can be removed without loss. Experiments across multiple attack types report extraction accuracy above 0.90 in most cases, exceeding previous methods and preserving diagnostic usability.

Core claim

Vol-Mark is a reversible-zero watermarking approach for medical volume data that designs a contrastive learning-based volume data feature extractor to obtain discriminative and stable volumetric features robust to 3D attacks, and employs cubic difference expansion using 3D integer wavelet transform to embed watermark bits into low-frequency coefficients within cubes by expanding voxel differences, using majority voting for reliable extraction, achieving low distortion and lossless removal while enabling integrity and ownership verification.

What carries the argument

Contrastive learning feature extractor paired with cubic difference expansion (c-DE) that uses 3D integer wavelet low-frequency coefficients inside voxel cubes to create embedding space and applies majority voting at extraction.

If this is right

  • Medical volumes can be shared across networks with both tampering detection and ownership proof before any diagnostic use.
  • The watermark can be stripped completely after verification so the original data remains available for repeated clinical analysis.
  • Dual-stage checking first tests integrity and then runs hypothesis-based ownership verification to limit errors when data has been altered.
  • The method maintains extraction accuracy above 0.90 across the tested attack categories while keeping embedding distortion low enough to avoid affecting diagnostic value.

Where Pith is reading between the lines

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

  • If the learned features transfer to other 3D imaging domains, the same pipeline could protect non-medical volumetric datasets such as those used in scientific visualization.
  • Embedding the mark at low-frequency wavelet coefficients may allow the technique to survive compression steps that are common in hospital data pipelines.
  • Combining the contrastive extractor with downstream diagnostic models could let ownership checks run automatically whenever a volume is loaded for AI-assisted analysis.

Load-bearing premise

The contrastive learning model yields volumetric features that remain sufficiently stable and distinctive after the full range of conventional, geometric, and hybrid attacks that might occur during data sharing.

What would settle it

A set of tests on rotated or scaled volumes where the ownership verification accuracy drops below 0.80 while the data still passes visual diagnostic checks.

Figures

Figures reproduced from arXiv: 2605.04705 by Jiangnan Zhu, Shengli Pan, Yujie Gu, Yuntao Wang.

Figure 1
Figure 1. Figure 1: Medical volume data. Watermarking is a widely used technology that enables healthcare institutions to securely store and transmit medical volume data while protecting it against unauthorized access and cyberattacks [41]. Current watermarking methods for med￾ical images can be broadly categorized into three types: region￾of-interest (ROI) lossless watermarking, reversible watermark￾ing, and zero-watermarkin… view at source ↗
Figure 2
Figure 2. Figure 2: The workflow of our proposed Vol-mark method. (a) First, Vol-Mark extracts features from volume data using a view at source ↗
Figure 3
Figure 3. Figure 3: 3D integer wavelet transform scheme. IV. PROPOSED METHOD In this section, we present our Vol-Mark method. It con￾sists of three phases: watermark registration, extraction and recovery, and verification. A. Watermark Registration Watermark registration phase consists of three processes: feature extraction, ownership share generation and embedding. 1) Feature extraction for volume data: Existing transform￾ba… view at source ↗
Figure 4
Figure 4. Figure 4: Architecture of the feature extractor. (ii) Preprocessing. All input volumes are resized to the same resolution of 128 × 128 × 64. The data are then normalized using the global mean µ and standard deviation σ computed from the training dataset via Vnorm = V −µ σ . Data augmentation is applied to the original medical vol￾umes to generate positive pairs for contrastive learning. Specif￾ically, each volume is… view at source ↗
Figure 5
Figure 5. Figure 5: Cubic difference expansion (c-DE) process is fully based on integer operations. It employs 3D￾IWT, whose reversibility guarantees accurate data extraction without any loss. All other operations in c-DE also integer￾based, which further ensures lossless reconstruction and pre￾serves the reversibility of the overall embedding process. The workflow of c-DE for embedding OS into volume data is summarized in Al… view at source ↗
Figure 6
Figure 6. Figure 6: The slices and 3D models of samples in MSD dataset. view at source ↗
Figure 7
Figure 7. Figure 7: Visualization of volume data from Task01 (Brain Tumours) after conventional attacks. view at source ↗
Figure 8
Figure 8. Figure 8: Visualization of volume data from Task01 (Brain Tumours) after geometric attacks. view at source ↗
Figure 9
Figure 9. Figure 9: Visualization of volume data after random cropping attacks (5%). view at source ↗
Figure 10
Figure 10. Figure 10: Accuracy comparison of Vol-Mark and baselines under hybrid attacks. view at source ↗
Figure 11
Figure 11. Figure 11: Accuracy comparison of Vol-Mark and baselines under 3D attacks. view at source ↗
Figure 12
Figure 12. Figure 12: Comparison of accuracy under random cropping. view at source ↗
read the original abstract

Today, advances in medical technology extensively utilize 3D volume data for accurate and efficient diagnostics. However, sharing these data across networks in telemedicine poses significant security risks of data tampering and unauthorized copying. To address these challenges, this paper proposes a novel reversible-zero watermarking approach, termed Vol-Mark, for medical volume data to protect their ownership and authenticity in telemedicine. The proposed Vol-Mark method offers two key benefits: 1) it designs a volume data feature extractor that leverages contrastive learning to efficiently extract discriminative and stable volumetric features, ensuring robustness against 3D attacks; 2) it introduces the cubic difference expansion (c-DE) technique, which leverages the 3D integer wavelet transform to embed watermark bits into neighboring voxels within cubes at low-frequency coefficients. The voxel differences within each cube are expanded to create embedding space, and a majority voting mechanism is employed during extraction to enhance reliability. The embedding process incurs low distortion and supports lossless removal, thereby preserving the integrity and diagnostic accuracy of medical volume data. Through these two benefits, Vol-Mark enables both integrity verification and ownership verification. Integrity verification is first performed, and ownership verification through hypothesis testing is further conducted to enhance reliability, particularly under data tampering or watermark removal attacks. Comprehensive experimental results show the effectiveness of the proposed method and its superior robustness against conventional, geometric, and hybrid attacks on medical volume data. In particular, through multiple tasks evaluations, Vol-Mark consistently achieves an ACC above 0.90 in most attack scenarios, outperforming existing methods by a clear margin.

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

3 major / 2 minor

Summary. The paper proposes Vol-Mark, a reversible zero-watermarking scheme for 3D medical volume data. It extracts robust volumetric features via contrastive learning and embeds watermark bits using cubic difference expansion (c-DE) based on 3D integer wavelet transform applied to voxel cubes at low-frequency coefficients, with majority voting during extraction to improve reliability. The method supports integrity verification followed by ownership verification via hypothesis testing, claims low embedding distortion with full reversibility, and reports ACC above 0.90 under conventional, geometric, and hybrid attacks while outperforming prior methods.

Significance. If the robustness claims hold under detailed scrutiny, the work could meaningfully advance secure sharing of 3D medical imaging data in telemedicine by providing a reversible, low-distortion watermark that preserves diagnostic quality. The integration of contrastive learning for 3D feature stability with c-DE for embedding is a plausible technical contribution to medical data security.

major comments (3)
  1. [§3.2] §3.2 (Contrastive Feature Extractor): The description of the contrastive learning pipeline does not specify the 3D-specific augmentations (e.g., rotation ranges, scaling factors, or intensity perturbations), network backbone, or loss function hyperparameters. Without these, it is impossible to assess whether the claimed invariance of volumetric features to geometric attacks is actually achieved or merely assumed.
  2. [§4.2] §4.2 (Cubic Difference Expansion): The embedding equations for c-DE via 3D integer wavelet transform are presented at a high level, but the precise selection of low-frequency coefficients, the expansion factor, and the cube partitioning strategy lack explicit formulas or pseudocode. This leaves open whether the majority-voting extraction reliably recovers bits under noise or tampering without introducing diagnostic artifacts.
  3. [§5] §5 (Experimental Evaluation): The reported ACC > 0.90 across attack scenarios is not accompanied by dataset sizes, number of volumes per dataset, statistical tests (e.g., standard deviation over runs), or full attack parameterizations (e.g., exact rotation angles or compression ratios). This makes it difficult to verify the superiority margin over baselines or to rule out limited attack coverage.
minor comments (2)
  1. [Abstract] The abstract refers to 'multiple tasks evaluations' without defining them; this should be clarified in §1 or §5.1.
  2. [Figures] Figure captions for attack robustness plots should explicitly list the attack parameters used rather than generic labels.

Simulated Author's Rebuttal

3 responses · 0 unresolved

We thank the referee for the constructive and detailed feedback. We address each major comment below and will incorporate the requested clarifications and additions into the revised manuscript.

read point-by-point responses

  1. Referee: [§3.2] §3.2 (Contrastive Feature Extractor): The description of the contrastive learning pipeline does not specify the 3D-specific augmentations (e.g., rotation ranges, scaling factors, or intensity perturbations), network backbone, or loss function hyperparameters. Without these, it is impossible to assess whether the claimed invariance of volumetric features to geometric attacks is actually achieved or merely assumed.

    Authors: We agree that the current description of the contrastive feature extractor lacks sufficient implementation details. In the revised §3.2 we will explicitly list the 3D augmentations employed to promote invariance (rotation ranges, scaling factors, and intensity perturbations), the network backbone architecture, and the loss function together with its hyperparameters. These additions will allow readers to evaluate how the extracted features achieve the reported robustness against geometric attacks. revision: yes

  2. Referee: [§4.2] §4.2 (Cubic Difference Expansion): The embedding equations for c-DE via 3D integer wavelet transform are presented at a high level, but the precise selection of low-frequency coefficients, the expansion factor, and the cube partitioning strategy lack explicit formulas or pseudocode. This leaves open whether the majority-voting extraction reliably recovers bits under noise or tampering without introducing diagnostic artifacts.

    Authors: The referee is correct that the c-DE description remains high-level. In the revision we will supply the exact formulas for low-frequency coefficient selection after the 3D integer wavelet transform, the expansion factor value, the cube partitioning strategy, and pseudocode for both the embedding and the majority-voting extraction steps. These additions will clarify how reversibility is preserved and how diagnostic quality is maintained. revision: yes

  3. Referee: [§5] §5 (Experimental Evaluation): The reported ACC > 0.90 across attack scenarios is not accompanied by dataset sizes, number of volumes per dataset, statistical tests (e.g., standard deviation over runs), or full attack parameterizations (e.g., exact rotation angles or compression ratios). This makes it difficult to verify the superiority margin over baselines or to rule out limited attack coverage.

    Authors: We acknowledge that the experimental section would benefit from greater transparency. The revised §5 will report the dataset sizes and number of volumes used, include standard deviations across repeated runs, and provide complete parameterizations for every attack (rotation angles, compression ratios, etc.). Statistical significance tests comparing Vol-Mark against baselines will also be added to strengthen the robustness claims. revision: yes

Circularity Check

0 steps flagged

No circularity in derivation chain

full rationale

The paper proposes Vol-Mark by combining a contrastive-learning feature extractor with cubic difference expansion (via 3D integer wavelet transform on cubes) plus majority voting. No equations, derivations, or load-bearing steps are shown that reduce any claimed prediction or result to a fitted parameter or self-referential definition by construction. The robustness claims rest on experimental outcomes rather than tautological mappings, and no self-citation chains or ansatzes imported from prior author work are invoked to force the central result. The derivation chain is therefore self-contained against external benchmarks.

Axiom & Free-Parameter Ledger

0 free parameters · 0 axioms · 0 invented entities

Abstract provides no explicit technical details on parameters, axioms, or new entities; the method builds on established contrastive learning and integer wavelet transforms without introducing or specifying any free parameters, axioms, or invented entities.

pith-pipeline@v0.9.0 · 5591 in / 1326 out tokens · 122779 ms · 2026-05-08T17:56:08.458645+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

51 extracted references · 2 canonical work pages

  1. [1]

    Watermarking techniques for medical data authentication: a survey,

    A. Anand and A. K. Singh, “Watermarking techniques for medical data authentication: a survey,”Multimedia Tools and Applications, vol. 80, no. 20, pp. 30 165–30 197, 2021

    2021

  2. [2]

    A hybrid optimization-based medical data hiding scheme for industrial internet of things security,

    ——, “A hybrid optimization-based medical data hiding scheme for industrial internet of things security,”IEEE Transactions on Industrial Informatics, vol. 19, no. 1, pp. 1051–1058, 2022

    2022

  3. [3]

    Protection of medical images and patient related information in healthcare: Using an intelligent and reversible watermarking technique,

    M. Arsalan, A. S. Qureshi, A. Khan, and M. Rajarajan, “Protection of medical images and patient related information in healthcare: Using an intelligent and reversible watermarking technique,”Applied Soft Computing, vol. 51, pp. 168–179, 2017

    2017

  4. [4]

    Watermarking protocol inspired kidney stone segmentation in iomt,

    P. V . B. Bayari, N. Tomar, G. Bhatnagar, and C. Chattopadhyay, “Watermarking protocol inspired kidney stone segmentation in iomt,” IEEE Journal of Biomedical and Health Informatics, vol. 30, no. 2, pp. 828–838, 2026

    2026

  5. [5]

    Multimodal face recognition method with two-dimensional hidden markov model,

    J. Bobulski, “Multimodal face recognition method with two-dimensional hidden markov model,”Bulletin of the Polish Academy of Sciences. Technical Sciences, vol. 65, no. 1, pp. 121–128, 2017

    2017

  6. [6]

    High capacity and reversible fragile watermarking method for medical image authentication and patient data hiding,

    R. Bouarroudj, F. Z. Bellala, and F. Souami, “High capacity and reversible fragile watermarking method for medical image authentication and patient data hiding,”Journal of Medical Systems, vol. 48, no. 1, p. 98, 2024

    2024

  7. [7]

    Wavelet transforms that map integers to integers,

    A. R. Calderbank, I. Daubechies, W. Sweldens, and B.-L. Yeo, “Wavelet transforms that map integers to integers,”Applied and Computational Harmonic Analysis, vol. 5, no. 3, pp. 332–369, 1998

    1998

  8. [8]

    A novel convolutional neural network model based on beetle antennae search optimization algorithm for computerized tomography diagnosis,

    D. Chen, X. Li, and S. Li, “A novel convolutional neural network model based on beetle antennae search optimization algorithm for computerized tomography diagnosis,”IEEE Transactions on Neural Networks and Learning Systems, vol. 34, no. 3, pp. 1418–1429, 2023

    2023

  9. [9]

    Med3d: Transfer learning for 3d medical image analysis

    S. Chen, K. Ma, and Y . Zheng, “Med3d: Transfer learning for 3d medical image analysis,”arXiv preprint arXiv:1904.00625, 2019

    work page arXiv 1904

  10. [10]

    A simple framework for contrastive learning of visual representations,

    T. Chen, S. Kornblith, M. Norouzi, and G. Hinton, “A simple framework for contrastive learning of visual representations,” inProceedings of the 37th International Conference on Machine Learning, ser. Proceedings of Machine Learning Research, H. D. III and A. Singh, Eds., vol. 119. PMLR, 13–18 Jul 2020, pp. 1597–1607

    2020

  11. [11]

    Adaptive region-specific loss for improved medical image segmentation,

    Y . Chen, L. Yu, J.-Y . Wang, N. Panjwani, J.-P. Obeid, W. Liu, L. Liu, N. Kovalchuk, M. F. Gensheimer, L. K. Vitzthum, B. M. Beadle, D. T. Chang, Q.-T. Le, B. Han, and L. Xing, “Adaptive region-specific loss for improved medical image segmentation,”IEEE Transactions on Pattern Analysis and Machine Intelligence, vol. 45, no. 11, pp. 13 408–13 421, 2023

    2023

  12. [12]

    Chapter 2 - applications and properties,

    I. J. Cox, M. L. Miller, J. A. Bloom, J. Fridrich, and T. Kalker, “Chapter 2 - applications and properties,” inDigital Watermarking and Steganography, 2nd ed. Burlington: Morgan Kaufmann, 2008, pp. 15– 59

    2008

  13. [13]

    Scalable watermarking for identifying large language model outputs,

    S. Dathathri, A. See, S. Ghaisas, P.-S. Huang, R. McAdam, J. Welbl, V . Bachani, A. Kaskasoli, R. Stanforth, T. Matejovicovaet al., “Scalable watermarking for identifying large language model outputs,”Nature, vol. 634, no. 8035, pp. 818–823, 2024

    2024

  14. [14]

    Automatic searching and pruning of deep neural networks for medical imaging diagnostic,

    F. E. Fernandes and G. G. Yen, “Automatic searching and pruning of deep neural networks for medical imaging diagnostic,”IEEE Transac- tions on Neural Networks and Learning Systems, vol. 32, no. 12, pp. 5664–5674, 2021

    2021

  15. [15]

    Efficient robust reversible watermarking based on zms and integer wavelet transform,

    G. Gao, M. Wang, and B. Wu, “Efficient robust reversible watermarking based on zms and integer wavelet transform,”IEEE Transactions on Industrial Informatics, vol. 20, no. 3, pp. 4115–4123, 2023. 15

    2023

  16. [16]

    Application of robust zero-watermarking scheme based on federated learning for securing the healthcare data,

    B. Han, R. H. Jhaveri, H. Wang, D. Qiao, and J. Du, “Application of robust zero-watermarking scheme based on federated learning for securing the healthcare data,”IEEE Journal of Biomedical and Health Informatics, vol. 27, no. 2, pp. 804–813, 2023

    2023

  17. [17]

    A new robust zero-watermarking algorithm for medical volume data,

    B. Han, J. Li, and L. Zong, “A new robust zero-watermarking algorithm for medical volume data,”International Journal of Signal Processing, Image Processing and Pattern Recognition, vol. 6, no. 6, pp. 245–258, 2013

    2013

  18. [18]

    Learning better registration to learn better few-shot medical image seg- mentation: Authenticity, diversity, and robustness,

    Y . He, R. Ge, X. Qi, Y . Chen, J. Wu, J.-L. Coatrieux, G. Yang, and S. Li, “Learning better registration to learn better few-shot medical image seg- mentation: Authenticity, diversity, and robustness,”IEEE Transactions on Neural Networks and Learning Systems, vol. 35, no. 2, pp. 2588– 2601, 2024

    2024

  19. [19]

    A two-dimensional mapping with a strange attractor,

    M. H ´enon, “A two-dimensional mapping with a strange attractor,” Communications in Mathematical Physics, vol. 50, no. 1, pp. 69–77, 1976

    1976

  20. [20]

    Arwgan: Attention-guided robust image watermarking model based on gan,

    J. Huang, T. Luo, L. Li, G. Yang, H. Xu, and C.-C. Chang, “Arwgan: Attention-guided robust image watermarking model based on gan,”IEEE Transactions on Instrumentation and Measurement, vol. 72, pp. 1–17, 2023

    2023

  21. [21]

    Boosting memory efficiency in transfer learning for high-resolution medical image classification,

    Y . Huang, P. Cheng, R. Tam, and X. Tang, “Boosting memory efficiency in transfer learning for high-resolution medical image classification,” IEEE Transactions on Neural Networks and Learning Systems, vol. 36, no. 9, pp. 17 280–17 294, 2025

    2025

  22. [22]

    Medical transformer: Universal encoder for 3-d brain mri analysis,

    E. Jun, S. Jeong, D.-W. Heo, and H.-I. Suk, “Medical transformer: Universal encoder for 3-d brain mri analysis,”IEEE Transactions on Neural Networks and Learning Systems, vol. 35, no. 12, pp. 17 779– 17 789, 2024

    2024

  23. [23]

    Logistic chaotic maps for binary numbers generations,

    A. Kanso and N. Smaoui, “Logistic chaotic maps for binary numbers generations,”Chaos, Solitons & Fractals, vol. 40, no. 5, pp. 2557–2568, 2009

    2009

  24. [24]

    3-d quantum-inspired self-supervised tensor network for volumetric segmentation of medical images,

    D. Konar, S. Bhattacharyya, T. K. Gandhi, B. K. Panigrahi, and R. Jiang, “3-d quantum-inspired self-supervised tensor network for volumetric segmentation of medical images,”IEEE Transactions on Neural Net- works and Learning Systems, vol. 35, no. 8, pp. 10 312–10 325, 2024

    2024

  25. [25]

    Identifying appropriate intellectual property protection mechanisms for machine learning models: A system- atization of watermarking, fingerprinting, model access, and attacks,

    I. Lederer, R. Mayer, and A. Rauber, “Identifying appropriate intellectual property protection mechanisms for machine learning models: A system- atization of watermarking, fingerprinting, model access, and attacks,” IEEE Transactions on Neural Networks and Learning Systems, vol. 35, no. 10, pp. 13 082–13 100, 2024

    2024

  26. [26]

    Geometric attacks on image watermarking systems,

    V . Licks and R. Jordan, “Geometric attacks on image watermarking systems,”IEEE Multimedia, vol. 12, no. 3, pp. 68–78, 2005

    2005

  27. [27]

    Robust watermark- ing algorithm for medical volume data in internet of medical things,

    J. Liu, J. Ma, J. Li, M. Huang, N. Sadiq, and Y . Ai, “Robust watermark- ing algorithm for medical volume data in internet of medical things,” IEEE Access, vol. 8, pp. 93 939–93 961, 2020

    2020

  28. [28]

    A novel zero-watermarking scheme with enhanced distin- guishability and robustness for volumetric medical imaging,

    X. Liu, Y . Sun, J. Wang, C. Yang, Y . Zhang, L. Wang, Y . Chen, and H. Fang, “A novel zero-watermarking scheme with enhanced distin- guishability and robustness for volumetric medical imaging,”Signal Processing: Image Communication, vol. 92, p. 116124, 2021

    2021

  29. [29]

    Attack-defending contrastive learning for volumetric medical image zero-watermarking,

    X. Liu, C. Yang, J. He, H. Fang, G. Schaefer, J. Zhang, Y . Zhu, and S. Zhang, “Attack-defending contrastive learning for volumetric medical image zero-watermarking,”ACM Transactions on Multimedia Computing, Communications and Applications, vol. 21, no. 2, pp. 1–23, 2025

    2025

  30. [30]

    Deep learning-based dual watermarking for image copyright protection and authentication,

    S. K. Padhi, A. Tiwari, and S. S. Ali, “Deep learning-based dual watermarking for image copyright protection and authentication,”IEEE Transactions on Artificial Intelligence, vol. 5, no. 12, pp. 6134–6145, 2024

    2024

  31. [31]

    Ai in medical imaging informatics: current challenges and future directions,

    A. S. Panayides, A. Amini, N. D. Filipovic, A. Sharma, S. A. Tsaftaris, A. Young, D. Foran, N. Do, S. Golemati, T. Kurcet al., “Ai in medical imaging informatics: current challenges and future directions,”IEEE Journal of Biomedical and Health Informatics, vol. 24, no. 7, pp. 1837– 1857, 2020

    2020

  32. [32]

    Watermarking deep neural networks in image processing,

    Y . Quan, H. Teng, Y . Chen, and H. Ji, “Watermarking deep neural networks in image processing,”IEEE Transactions on Neural Networks and Learning Systems, vol. 32, no. 5, pp. 1852–1865, 2021

    2021

  33. [33]

    A comprehensive survey of cybersecurity threats and data privacy issues in healthcare systems,

    R. Qureshi and I. Koo, “A comprehensive survey of cybersecurity threats and data privacy issues in healthcare systems,”Applied Sciences, vol. 16, no. 3, p. 1511, 2026

    2026

  34. [34]

    Roi-based medical image watermarking for accu- rate tamper detection, localisation and recovery,

    D. Ravichandran, P. Praveenkumar, S. Rajagopalan, J. B. B. Rayappan, and R. Amirtharajan, “Roi-based medical image watermarking for accu- rate tamper detection, localisation and recovery,”Medical & Biological Engineering & Computing, vol. 59, no. 6, pp. 1355–1372, 2021

    2021

  35. [35]

    A new approach to fully-reversible watermarking in medical imaging with breakthrough vis- ibility parameters,

    A. Ro ˇcek, K. Slav´ıˇcek, O. Dost ´al, and M. Javorn´ık, “A new approach to fully-reversible watermarking in medical imaging with breakthrough vis- ibility parameters,”Biomedical Signal Processing and Control, vol. 29, pp. 44–52, 2016

    2016

  36. [36]

    A large annotated medical image dataset for the development and evaluation of segmentation algorithms

    A. L. Simpson, M. Antonelli, S. Bakas, M. Bilello, K. Farahani, B. Van Ginneken, A. Kopp-Schneider, B. A. Landman, G. Litjens, B. Menzeet al., “A large annotated medical image dataset for the development and evaluation of segmentation algorithms,”arXiv preprint arXiv:1902.09063, 2019

    work page Pith review arXiv 1902

  37. [37]

    A reversible-zero watermarking scheme for medical images,

    R. Taj, F. Tao, S. Kanwal, A. Almogren, A. Altameem, and A. Ur Rehman, “A reversible-zero watermarking scheme for medical images,”Scientific Reports, vol. 14, no. 1, p. 17320, 2024

    2024

  38. [38]

    Robust image watermarking theories and techniques: A review,

    H. Tao, L. Chongmin, J. M. Zain, and A. N. Abdalla, “Robust image watermarking theories and techniques: A review,”Journal of Applied Research and Technology, vol. 12, no. 1, pp. 122–138, 2014

    2014

  39. [39]

    Reversible data embedding using a difference expansion,

    J. Tian, “Reversible data embedding using a difference expansion,”IEEE Transactions on Circuits and Systems for Video Technology, vol. 13, no. 8, pp. 890–896, 2003

    2003

  40. [40]

    High capacity reversible data hiding with interpolation and adaptive embedding,

    M. A. Wahed and H. Nyeem, “High capacity reversible data hiding with interpolation and adaptive embedding,”PloS One, vol. 14, no. 3, p. e0212093, 2019

    2019

  41. [41]

    Image copyright protec- tion based on blockchain and zero-watermark,

    B. Wang, S. Jiawei, W. Wang, and P. Zhao, “Image copyright protec- tion based on blockchain and zero-watermark,”IEEE Transactions on Network Science and Engineering, vol. 9, no. 4, pp. 2188–2199, 2022

    2022

  42. [42]

    Learning lossless compression for high bit-depth volumetric medical image,

    K. Wang, Y . Bai, D. Li, D. Zhai, J. Jiang, and X. Liu, “Learning lossless compression for high bit-depth volumetric medical image,” IEEE Transactions on Image Processing, vol. 34, pp. 113–125, 2025

    2025

  43. [43]

    Medical image segmentation using deep learning: A survey,

    R. Wang, T. Lei, R. Cui, B. Zhang, H. Meng, and A. K. Nandi, “Medical image segmentation using deep learning: A survey,”IET Image Processing, vol. 16, no. 5, pp. 1243–1267, 2022

    2022

  44. [44]

    Concept and application of zero- watermark,

    Q. Wen, T.-F. Sun, and S.-X. Wang, “Concept and application of zero- watermark,”Acta Electronica Sinica, vol. 31, no. 2, pp. 214–216, 2003

    2003

  45. [45]

    Role exchange-based self-training semi-supervision framework for complex medical image segmentation,

    Y . Wu, G. Wu, J. Lin, Y . Wang, and J. Yu, “Role exchange-based self-training semi-supervision framework for complex medical image segmentation,”IEEE Transactions on Neural Networks and Learning Systems, vol. 36, no. 5, pp. 8372–8386, 2025

    2025

  46. [46]

    A trusted medical image zero-watermarking scheme based on dcnn and hyperchaotic system,

    R. Xiang, G. Liu, M. Dang, Q. Wang, and R. Pan, “A trusted medical image zero-watermarking scheme based on dcnn and hyperchaotic system,”IEEE Journal of Biomedical and Health Informatics, vol. 29, no. 6, pp. 4241–4253, 2025

    2025

  47. [47]

    A multiwatermarking scheme for verifying medical image integrity and authenticity in the internet of medical things,

    F. Yan, H. Huang, and X. Yu, “A multiwatermarking scheme for verifying medical image integrity and authenticity in the internet of medical things,”IEEE Transactions on Industrial Informatics, vol. 18, no. 12, pp. 8885–8894, 2022

    2022

  48. [48]

    Robust multi-watermarking algorithm for medical images based on googlenet and henon map

    W. Zhang, J. Li, U. A. Bhatti, J. Liu, J. Zheng, and Y .-W. Chen, “Robust multi-watermarking algorithm for medical images based on googlenet and henon map.”Computers, Materials & Continua, vol. 75, no. 1, 2023

    2023

  49. [49]

    Diagnose like a radiologist: Hybrid neuro-probabilistic reasoning for attribute-based medical image diagnosis,

    G. Zhao, Q. Feng, C. Chen, Z. Zhou, and Y . Yu, “Diagnose like a radiologist: Hybrid neuro-probabilistic reasoning for attribute-based medical image diagnosis,”IEEE Transactions on Pattern Analysis and Machine Intelligence, vol. 44, no. 11, pp. 7400–7416, 2022

    2022

  50. [50]

    nn- former: V olumetric medical image segmentation via a 3d transformer,

    H.-Y . Zhou, J. Guo, Y . Zhang, X. Han, L. Yu, L. Wang, and Y . Yu, “nn- former: V olumetric medical image segmentation via a 3d transformer,” IEEE Transactions on Image Processing, vol. 32, pp. 4036–4045, 2023

    2023

  51. [51]

    A reversible-zero watermarking scheme for medical volume data via difference expansion,

    J. Zhu, Y . Wang, and Y . Gu, “A reversible-zero watermarking scheme for medical volume data via difference expansion,” inProceedings of the 2025 IEEE Conference on Artificial Intelligence. IEEE, 2025, pp. 420–423

    2025