Pith. sign in

REVIEW 5 major objections 6 minor 54 references

Efficient and Accurate Image Provenance Analysis: A Scalable Pipeline for Large-scale Images

T0 review · 5 major / 6 minor · reviewed 2026-08-06 · deepseek-v4-flash

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

Pith's one-line read The paper claims that tracing pre-maintained modification links turns image provenance analysis into a linear-time problem, letting a pipeline recover heavily edited variants of a query in about 3 seconds on 10-million-image databases.

desk verdict The MR tracing idea is genuinely new, but the validation is circular: the MR store appears seeded with ground-truth edges, and the headline accuracy and speed gains all come from that lookup. read the letter →

arxiv 2506.23707 v1 pith:B64IC6AY submitted 2025-06-30 cs.MM

Jiewei Lai , Lan Zhang , Chen Tang , Pengcheng Sun This is my paper · ORCID
classification cs.MM
keywords imageprovenanceanalysisdigitalforensicsgraphconstructionmodificationrelationshiptracingJPEGcompressionartifactslarge-scaleretrievalmultimediaphylogenysimilarityfiltering
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

Image provenance analysis asks two things of a query image: find every variant in a large database that descends from it through edits, and reconstruct the directed family tree of those edits. Previous systems compare images pairwise, which both misses heavily modified copies whose similarity to the original is low and costs $O(n^2)$ time, making ten-million-image databases impractical. This paper argues that both weaknesses can be removed at once if the database already maintains its modification relationships: filtering can walk those links to pull in low-similarity variants, and graph construction can read the links instead of re-analyzing candidate pairs. On that basis the paper builds an end-to-end pipeline whose per-query cost is $O(n)$, reporting a 16.7–56.1% accuracy improvement in end-to-end provenance analysis and an average 3.0-second response at 10-million scale, versus about 12 minutes for the prior state of the art. The whole argument therefore rests on one premise: that a correct, complete modification-relationship graph already exists inside the database.

What carries the argument

The machinery is threefold. First, MR tracing: after an ordinary top-$k$ similarity query, the pipeline walks the directed graph of modification relationships that the database already maintains, returning the ancestors and descendants of each discovered node; a four-case analysis of three-image configurations justifies returning exactly those images, since only they share a modification relationship with the query. Second, an MR analytical network, a Siamese architecture that fuses hierarchical feature maps of two images and classifies whether they share direct content, establishing undirected edges and discarding unrelated images where minimum-spanning-tree methods could not. Third, a direction determination network that, for each undirected edge, takes the residuals of both images under a learned JPEG-artifact-reduction model and classifies the direction; the operative signal is that a re-compressed modified image exhibits dual artifacts ($f_0$ and $f_1$) while its source exhibits only $f_0$.

What would settle it

Run the pipeline on a database whose MR store must be built from scratch, timing store construction separately: if populating the store from a 10-million-image database requires pairwise analysis of all candidate pairs, the total wall-clock cost grows quadratically and the linear-time claim holds only for queries on a pre-built store. A second check: take a provenance chain whose images were never JPEG-compressed (for example, PNG or TIFF only) and test the direction network; if accuracy on such pairs falls to chance, the dual-artifact mechanism is confirmed as the operative signal and the method's stated robustness is bounded to JPEG-family pipelines.

Watch

Extended reading notes

Core claim

The paper's central claim is that modification relationships (MR) among images in a database are reusable assets rather than something to be recomputed per query. Once those relationships are recorded as a directed graph, a query's provenance can be recovered by (1) an MR tracing step that returns the ancestors and descendants of any top-$k$ candidate, so that variants with arbitrarily low visual similarity to the query are still discovered, and (2) a graph construction step that only analyzes the query against its candidates, fetching the relations among candidates from the stored graph. The paper further claims that modification direction can be read from JPEG compression history: a modified image carries dual compression artifacts, the original artifact $f_0$ of its source plus a new artifact $f_1$ introduced by re-compression, while the source carries only $f_0$; residual images produced by an artifact-reduction model expose this difference to a learned direction classifier. Combined with a global image representation that makes similarity scoring $O(1)$, these steps give the pipeline a claimed per-analysis complexity of $O(n)$ and yield empirical accuracy gains over prior methods on three public provenance datasets.

Load-bearing premise

The pipeline's linear-time speed and its recovery of low-similarity variants rest on a single premise: that a complete, correct directed graph of modification relationships among the database's images already exists and is maintained by the database, so tracing it costs nothing and adds no errors.

Editorial extensions

If this is right

  • If the pipeline works as claimed, ten-million-image databases become practically searchable for provenance: roughly 3 seconds per query end-to-end instead of the 12-minute graph-construction cost of the prior approach.
  • Heavily modified images that rank far outside the top-$k$ similarity candidates are still recovered, since MR tracing pulls in the full ancestor-descendant set of any candidate that is found.
  • The same MR tracing step makes graph construction linear: the pipeline analyzes only the query–candidate pairs (at most $n-1$) and reads candidate-to-candidate relations from the store, so total per-query complexity is $O(n)$ rather than $O(n^2)$.
  • The JPEG dual-artifact signal gives a direction signal that does not depend on visible visual differences, which the paper reports at 85.3% average accuracy across 21 modification scenarios.
  • MR tracing is a plug-in: the paper applies it to four different filtering baselines and reports consistent recall gains, so the benefit is not tied to one particular image representation.

Reading between the lines

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

  • The reported experiments appear to populate the MR store with the datasets' ground-truth provenance edges, so the paper does not measure the cost or error of building that store automatically; a real deployment would inherit both the construction cost and any errors in the store.
  • If the MR store must be maintained, the quadratic work the paper eliminates from queries reappears at ingestion time unless relationships are derived from edit metadata, watermarks, or content credentials when images enter the database.
  • Because MR tracing returns the whole ancestor-descendant closure of any discovered node, a single spurious edge in the stored graph propagates directly into query results; the pipeline's precision is upper-bounded by the quality of the maintained graph.
  • The dual-artifact signal is JPEG-specific, and the paper's own results on mixed-format datasets show the direction model degrading; testing the residual-difference idea on WebP/AVIF or lossless re-encoding chains would be a natural next experiment.
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

5 major / 6 minor

Summary. The paper proposes an end-to-end image provenance analysis pipeline. It combines conventional top-k similarity filtering with an MR-tracing step that reads modification relationships from a graph database, then constructs a directed provenance graph using a learned local-feature matching model for edge existence and a JPEG-artifact model for edge direction. The authors claim O(n) end-to-end complexity, an average 3.0-second response time at ten-million scale, and a 16.7–56.1% accuracy improvement over prior methods. The central mechanism of the pipeline is that relationships among candidate images are not computed during query time but are retrieved from a pre-maintained MR store.

Significance. If the MR store could be constructed independently and inexpensively, the proposed pipeline would be a meaningful practical contribution to large-scale image provenance analysis: it combines a global-representation retrieval stage with learned pairwise analysis and has a clear engineering story for avoiding repeated pairwise computation. The paper also provides useful detail: complete pseudocode, experiments on three provenance benchmarks, an ablation of the MR-tracing step, and robustness tests over 22 modification scenarios. These strengths are real and should be credited. However, the end-to-end accuracy and complexity claims are not supported as stated because they depend on a pre-existing MR store whose construction is never described and that, on the benchmark datasets, appears to be populated with ground-truth provenance edges. The main novelty therefore reduces to a graph-database lookup layer, and the reported improvements do not establish that the pipeline discovers unknown provenance relationships.

major comments (5)
  1. [Sec. 4.2, Algorithm 1, Algorithm 3, Table 4] The MR-tracing step retrieves candidate images and edges directly from a pre-maintained MR store, but the paper never specifies how that store is built in the experiments. Section 4.2 states that 'such an MR can be known or pre-analyzed by the database maintainer' and Algorithm 3 (lines 18–29) reads edges out of the store rather than computing them. All three benchmarks come with ground-truth provenance graphs, and the only reported MR-store statistics (Section 5.6.2: approximately 12,000 MR, 0.66 MB, fixed 0.027s tracing time) are consistent with injecting the reference edges. If that is what was done, the recall gains in Table 2 and the VO/EO/VEO gains in Table 4 are obtained by looking up known provenance, making the headline '16.7–56.1% accuracy improvement' circular. The authors must either report a non-trivial procedure for building the MR store without using ground truth, or clearly state that the store is oracle-provided and restrict their claims accordingly.
  2. [Sec. 4.4, Complexity Analysis] The claimed O(n) complexity is not end-to-end because it excludes the cost of constructing and maintaining the MR store. The analysis counts only n−1 query–candidate relationship analyses and a fixed-cost MR lookup, but the MR store itself must be built by some process. If that process is pairwise image analysis, the quadratic cost is merely moved offline rather than eliminated; if the store is externally supplied, the comparison with end-to-end SOTA methods (IPA, GEVT) in Table 4 is not apples-to-apples. The complexity statement should either include MR-store construction and update costs or be explicitly scoped to query-time cost under the assumption of a precomputed oracle.
  3. [Sec. 4.2, Algorithm 1] The claim that MR tracing discovers images 'regardless of their similarity level' is too strong. Algorithm 1 only adds images that share an MR with one of the top-k similarity candidates; it cannot discover a true variant of the query that is not connected to any top-k candidate by a stored MR edge. Moreover, if the query image itself is not a node in the MR store, the tracing step has no starting point. The method therefore depends on the completeness and currency of the MR store in a way that is not acknowledged in Sections 1 and 4.2. A precise statement of the conditions under which all true variants are recovered is needed.
  4. [Algorithm 2, lines 5–20] The graph-construction algorithm keeps at most one edge per node: for each image I_i, it selects a single partner with the highest MR confidence and adds one edge. This cannot represent multiple direct modification relationships, e.g., composite or spliced images with more than one parent, which the paper itself lists as a target scenario in Section 2.1 (references [11, 12, 38]). The discrepancy between the algorithm's behavior and the stated goal of mining 'all MR within an image set' (Section 3) should be resolved, and Table 3 should be interpreted in light of this limitation.
  5. [Eq. (1), Section 3] The formalization of provenance filtering is ill-posed as written. The objective is an unconstrained arg max over D_candidate of the intersection size with the ground-truth MR set; without a cardinality or precision constraint, the trivial solution D_candidate = D maximizes the objective. The optimization should be reformulated, for example with a budget on |D_candidate| or with recall/precision trade-off, before it can support the subsequent filtering discussion.
minor comments (6)
  1. [Abstract vs. Sec. 5.6.2] The abstract and introduction report an 'average 3.0-second response time,' while Section 5.6.2 states 'the total per-analysis time of approximately 2.3s even at the 10M scale.' The numbers should be reconciled.
  2. [Sec. 4.4] The statement that the ISC representation 'enables O(1) similarity calculations' is imprecise for a 10-million-image database; approximate nearest-neighbor search has its own index-based cost. The paper should report the retrieval cost more carefully, especially since Figure 4 shows filtering time growing from 0.107s to 0.220s as the scale increases.
  3. [Sec. 5.7.4] The text says 'More detailed robustness evaluation experiments are conducted in Section 4.7,' but the robustness experiments are in Section 5.7.4; the cross-reference is incorrect.
  4. [Sec. 4.2] There is a typo: 'copy-cove' should be 'copy-move' in the discussion of modification techniques.
  5. [Sec. 5.5.3] There is a typo in 'under oracle and disturb mdoes' — 'mdoes' should be 'modes'.
  6. [Algorithm 2] The algorithm sets A[i][index] to 1 for an 'undirected edge' but later overwrites it with a directed value, and the matrix entry A[index][i] is never updated symmetrically. The notation should be clarified to distinguish the undirected adjacency representation from the directed one.

Circularity Check

3 steps flagged · score 8.0 of 10

Headline accuracy and O(n) gains reduce to reading a pre-maintained MR store; the experiments report numbers consistent with populating that store from the ground-truth provenance graphs, making the end-to-end results partly a readout of the input.

  1. self definitional [Sec. 4.2 'Modification Relationship Tracing for Enhanced Provenance Filtering'; Algorithm 3 (End-to-End Image Provenance Analysis Pipeline), lines 18-29]
    "it is feasible given that such an MR can be known or pre-analyzed by the database maintainer and subsequently maintained within the graph database ... MR tracing E_candidate = {e1,e2,...} <- F_f(D,V) ... Add edge to edges set Add(E, e(Ii,Ij))"

    The MR-tracing stage, advertised as discovering low-similarity variants, is a readout of the maintained MR store. Algorithm 3 obtains E_candidate directly from F_f(D,V), i.e., from the database's pre-maintained directed graph, and inserts those edges and their endpoints into the output graph without any pairwise verification. Section 4.2 explicitly states that this MR 'can be known or pre-analyzed by the database maintainer,' meaning the ground-truth provenance edges are inputs, not outputs. Consequently, the Table 2 recall gains ('regardless of similarity') and Table 4 edge-overlap gains measure the extent to which the pipeline copies the input store; the output provenance graph is, by construction, partly identical to the input MR graph.

  2. fitted input called prediction [Sec. 5.1 Table 1 (dataset statistics) vs Sec. 5.6.2 (Storage overhead)]
    "NC2017DEV 378 110075 18.06 4.69 13.37 / MFC2018DEV 159 18897 39.79 12.60 27.19 / Reddit 171 9959 53.68 44.24 9.44 ... Our analysis reveals that we maintain approximately 12,000 MR, which only requires 0.66 MB (i.e., 680 KB) of storage space."

    The only reported MR-store statistics (about 12,000 edges) closely match the sum of ground-truth direct-edge counts implied by Table 1: 378x4.69 + 159x12.60 + 171x44.24 is approximately 11,341. The datasets are ingested by extracting directed provenance graphs from the official journal files, so the MR store is populated with the same reference edges used to compute VO/EO/VEO. The end-to-end EO and VEO scores in Table 4 therefore partly report the input ground-truth graph back rather than measuring independently discovered provenance; the MR-tracing contribution is effectively an oracle lookup folded into the reported accuracy.

1 more flagged steps
  1. other [Sec. 4.4 'A Scalable End-to-End Pipeline for Provenance Analysis' (Complexity Analysis)]
    "by leveraging these MR maintained within the database, our pipeline is optimized to analyze solely the relationships between the query image and its candidates, thereby avoiding pairwise analysis during the graph construction phase."

    The O(n) complexity argument moves the quadratic pairwise analysis into the construction and maintenance of the MR store, which the paper assumes ('known or pre-analyzed') but never specifies or benchmarks. For any database that does not already contain the MR graph, building the store requires pairwise comparison of image pairs, which is exactly the O(n^2) cost the paper claims to eliminate. Thus the 'end-to-end linear time' claim is conditional on an unstated preprocessing step whose cost is the same as the complexity the method purports to remove; the reported 0.027s fixed MR-tracing time measures only the lookup, not the construction.

full rationale

The central novelty of the paper is MR tracing: Algorithm 3 fetches candidate edges from a database-maintained MR store and adds them to the output provenance graph, while Sec. 4.2 concedes that the MR 'can be known or pre-analyzed by the database maintainer.' This makes the headline 'regardless of similarity' filtering and the large EO/VEO improvements in the end-to-end table partly a readout of pre-existing ground-truth edges. The O(n) claim similarly depends on that store already existing, with the quadratic pairwise cost deferred to unmeasured store construction. The paper does contain independent components — the local-feature MR network and JPEG-artifact direction network are trained and evaluated on oracle/disturb graph construction — so not every result is circular. However, the end-to-end accuracy and scaling claims that define the paper's contribution reduce, by the paper's own algorithms and statements, to looking up an input graph. This is definitional circularity, not merely excessive self-citation: the output provenance graph is partly identical to the input MR store by construction. Score 8 reflects that the central end-to-end claim is forced by the assumed input, while the component-level graph construction retains independent content.

Assumptions & free parameters 0 free parameters · 5 assumptions · 0 invented entities

The pipeline rests on the availability of a pre-maintained MR graph; without it, MR tracing cannot run and the linear-complexity reduction disappears. The JPEG artifact model is standard but assumed; the O(1) filtering claim requires an unspecified index; and the benchmark ground truth is used both for training and, apparently, for populating the MR store.

assumptions (5)
  • domain assumption A database of modification relationships (MR) among images exists and is maintained by the database operator.
    Sec 4.2 states 'such an MR can be known or pre-analyzed by the database maintainer and subsequently maintained within the graph database.' The MR tracing step and the O(n) complexity both rely on this premise.
  • domain assumption A modified image carries dual JPEG compression artifacts while its source carries only the initial artifact.
    Sec 4.3.2 Equations (7)-(10) model original and modified JPEG images; this standard double-compression property motivates the direction determination network and is not independently verified in this paper.
  • domain assumption Images sharing modification relationships share detectable content at multiple levels of detail, so a learned local feature matcher can identify them.
    Sec 4.3.1 'Insight' paragraph. The MR analytical network is trained on the assumption that shared content implies modification relationship.
  • ad hoc to paper The image retrieval index built on the ISC global representation yields effectively O(1) top-k query time on a 10-million-image database.
    Sec 4.4 complexity analysis claims 'ISC representation for image representation enables O(1) similarity calculations', but no index structure is described and the filtering time in Fig. 4 grows sublinearly, not constant.
  • domain assumption Ground-truth provenance graphs in the benchmark datasets are correct and are used for training and for populating the MR store.
    Sec 5.1-5.3. The experiments split graphs into 70/30 train/validation and the MR store size (12,000 MR) matches the datasets' ground-truth edge counts.

how reviews work

0 comments
Cite this review

Pith. "Pith review of Efficient and Accurate Image Provenance Analysis: A Scalable Pipeline for Large-scale Images." pith.science (2026). https://pith.science/paper/B64IC6AY

@misc{pith2026250623707,
  author       = {Pith},
  title        = {Pith review of: Efficient and Accurate Image Provenance Analysis: A Scalable Pipeline for Large-scale Images},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/B64IC6AY}},
  note         = {Machine review of arXiv:2506.23707}
}
read the original abstract

The rapid proliferation of modified images on social networks that are driven by widely accessible editing tools demands robust forensic tools for digital governance. Image provenance analysis, which filters various query image variants and constructs a directed graph to trace their phylogeny history, has emerged as a critical solution. However, existing methods face two fundamental limitations: First, accuracy issues arise from overlooking heavily modified images due to low similarity while failing to exclude unrelated images and determine modification directions under diverse modification scenarios. Second, scalability bottlenecks stem from pairwise image analysis incurs quadratic complexity, hindering application in large-scale scenarios. This paper presents a scalable end-to-end pipeline for image provenance analysis that achieves high precision with linear complexity. This improves filtering effectiveness through modification relationship tracing, which enables the comprehensive discovery of image variants regardless of their visual similarity to the query. In addition, the proposed pipeline integrates local features matching and compression artifact capturing, enhancing robustness against diverse modifications and enabling accurate analysis of images' relationships. This allows the generation of a directed provenance graph that accurately characterizes the image's phylogeny history. Furthermore, by optimizing similarity calculations and eliminating redundant pairwise analysis during graph construction, the pipeline achieves a linear time complexity, ensuring its scalability for large-scale scenarios. Experiments demonstrate pipeline's superior performance, achieving a 16.7-56.1% accuracy improvement. Notably, it exhibits significant scalability with an average 3.0-second response time on 10 million scale images, which is far shorter than the SOTA approach's 12-minute duration.

Figures

Figures reproduced from arXiv: 2506.23707 by the authors.

Figure 1
Figure 1. Example of a provenance graph from the Reddit dataset, where nodes represent images and directed [PITH_FULL_IMAGE:figures/full_fig_p002_1.png] view at source ↗
Figure 2
Figure 2. The left panel shows four potential modification relationships for images [PITH_FULL_IMAGE:figures/full_fig_p007_2.png] view at source ↗
Figure 3
Figure 3. An overview of the proposed framework for constructing directed provenance graphs. The top panel [PITH_FULL_IMAGE:figures/full_fig_p009_3.png] view at source ↗
Figures from the paper (3 more)
Figure 4
Figure 4. Figure 4: Time cost of our image provenance analysis pipeline across different data scales. [PITH_FULL_IMAGE:figures/full_fig_p020_4.png]
Figure 5
Figure 5. Figure 5: Evaluation of MR analytical and direction determination Models. This figure presents the F1-scores [PITH_FULL_IMAGE:figures/full_fig_p021_5.png]
Figure 6
Figure 6. Figure 6: The evaluation results of our models in 22 different scenarios are presented. These include the tasks of [PITH_FULL_IMAGE:figures/full_fig_p022_6.png]

Discussion (0). Continue with ORCID to comment.

Reference graph

Works this paper leans on

54 extracted references · 33 canonical work pages

  1. [1]

    Maksym Andriushchenko, Xiaoyang Rebecca Li, Geoffrey Oxholm, Thomas Gittings, Tu Bui, Nicolas Flammarion, and John Collomosse. 2022. ARIA: Adversarially Robust Image Attribution for Content Provenance. In2022 IEEE/CVF Conference on Computer Vision and Pattern Recognition Workshops (CVPRW). IEEE, New Orleans, LA, USA, 33–43. https://doi.org/10.1109/CVPRW56...

    arXiv 2022

  2. [2]

    J Aythora, R Burke-Agüero, A Chamayou, S Clebsch, M Costa, J Deutscher, N Earnshaw, L Ellis, P England, C Fournet, et al. 2020. Multi-stakeholder media provenance management to counter synthetic media risks in news publishing. InProc. Intl. Broadcasting Convention (IBC), Vol. 1. 8. https://canucklaw.ca/wp-content/uploads/2021/04/Media- Provenance-Counteri...

    work page 2020

  3. [3]

    Lin Bai, Caiyan Jia, Ziying Song, and Chaoqun Cui. 2025. VGA: Vision and Graph Fused Attention Network for Rumor Detection.ACM Trans. Knowl. Discov. Data(March 2025). https://doi.org/10.1145/3722225 Just Accepted

    work page doi:10.1145/3722225 2025

  4. [4]

    Herbert Bay, Tinne Tuytelaars, and Luc Van Gool. 2006. SURF: Speeded Up Robust Features. InComputer Vision - ECCV 2006, 9th European Conference on Computer Vision, Graz, Austria, May 7-13, 2006, Proceedings, Part I (Lecture Notes in Computer Science, Vol. 3951), Ales Leonardis, Horst Bischof, and Axel Pinz (Eds.). Springer, 404–417. https: //doi.org/10.10...

    doi:10.1007/11744023_32 2006

  5. [5]

    Aparna Bharati, Daniel Moreira, Joel Brogan, Patricia Hale, Kevin Bowyer, Patrick Flynn, Anderson Rocha, and Walter Scheirer. 2019. Beyond Pixels: Image Provenance Analysis Leveraging Metadata. In2019 IEEE Winter Conference on Applications of Computer Vision (W ACV). IEEE, Waikoloa, HI, USA, 1692–1702. https://doi.org/10.1109/WACV.2019. 00185

    work page doi:10.1109/wacv.2019 2019

  6. [6]

    Flynn, Anderson de Rezende Rocha, Kevin W

    Aparna Bharati, Daniel Moreira, Patrick J. Flynn, Anderson de Rezende Rocha, Kevin W. Bowyer, and Walter J. Scheirer

    work page

  7. [7]

    Bharati, D

    A. Bharati, D. Moreira, A. Pinto, J. Brogan, K. Bowyer, P. Flynn, W. Scheirer, and A. Rocha. 2017. U-Phylogeny: Undirected provenance graph construction in the wild. In2017 IEEE International Conference on Image Processing (ICIP). IEEE, Beijing, China, 1517–1521. https://doi.org/10.1109/ICIP.2017.8296535

    work page arXiv 2017

  8. [8]

    Emil Bureaca and Iulian Aciobanitei. 2024. A Blockchain Blockchain-based Framework for Content Provenance and Authenticity. In16th International Conference on Electronics, Computers and Artificial Intelligence, ECAI 2024, Iasi, Romania, June 27-28, 2024. IEEE, 1–5. https://doi.org/10.1109/ECAI61503.2024.10607477

    arXiv 2024

Show all 54 references
  1. [9]

    Davis, and Ser-Nam Lim

    Bor-Chun Chen, Zuxuan Wu, Larry S. Davis, and Ser-Nam Lim. 2021. Efficient Object Embedding for Spliced Image Retrieval. In2021 IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR). IEEE, Nashville, TN, USA, 14960–14970. https://doi.org/10.1109/CVPR46437.2021.01472

    2021

  2. [10]

    Ting Chen, Simon Kornblith, Mohammad Norouzi, and Geoffrey Hinton. 2020. A Simple Framework for Contrastive Learning of Visual Representations. InProceedings of the 37th International Conference on Machine Learning (Proceedings of Machine Learning Research, Vol. 119), Hal Daum...

    2020

  3. [11]

    de Oliveira, Pasquale Ferrara, Alessia De Rosa, Alessandro Piva, Mauro Barni, Siome Goldenstein, Zanoni Dias, and Anderson Rocha

    Alberto A. de Oliveira, Pasquale Ferrara, Alessia De Rosa, Alessandro Piva, Mauro Barni, Siome Goldenstein, Zanoni Dias, and Anderson Rocha. 2016. Multiple Parenting Phylogeny Relationships in Digital Images.IEEE Transactions on Information Forensics and Security11, 2 (2016), ...

    2016

  4. [12]

    Zanoni Dias, Siome Goldenstein, and Anderson Rocha. 2013. Toward image phylogeny forests: Automatically recovering semantically similar image relationships.Forensic Science International231, 1 (2013), 178–189. https://doi.org/10.1016/ j.forsciint.2013.05.002

    2013

  5. [13]

    Zanoni Dias, Anderson Rocha, and Siome Goldenstein. 2010. First steps toward image phylogeny. In2010 IEEE International Workshop on Information Forensics and Security. IEEE, Seattle, WA, USA, 1–6. https://doi.org/10.1109/ WIFS.2010.5711452

    2010

  6. [14]

    Xiaojuan Dong, Weiming Zhang, Mohsin Shah, Bei Wang, and Nenghai Yu. 2020. Watermarking-based secure plaintext image protocols for storage, show, deletion and retrieval in the cloud.IEEE Transactions on Services Computing15, 3 (2020), 1678–1692. https://doi.org/10.1109/TSC.202...

    2020

  7. [15]

    Shanshan Feng, Guoxin Yu, Dawei Liu, Han Hu, Yong Luo, Hui Lin, and Yew-Soon Ong. 2025. MHR: A Multi-Modal Hyperbolic Representation Framework for Fake News Detection.IEEE Trans. Knowl. Data Eng.37, 4 (2025), 2015–2028. https://doi.org/10.1109/TKDE.2025.3528951

    2025

  8. [16]

    Zan Gao, Chao Sun, Zhiyong Cheng, Weili Guan, Anan Liu, and Meng Wang. 2023. TBNet: A Two-Stream Boundary- Aware Network for Generic Image Manipulation Localization.IEEE Trans. Knowl. Data Eng.35, 7 (2023), 7541–7556. https://doi.org/10.1109/TKDE.2022.3187091

    2023

  9. [17]

    Tiezheng Ge, Kaiming He, Qifa Ke, and Jian Sun. 2013. Optimized Product Quantization for Approximate Nearest Neighbor Search. In2013 IEEE Conference on Computer Vision and Pattern Recognition, Portland, OR, USA, June 23-28,

    2013

  10. [18]

    Haiying Guan, Andrew Delgado, Yooyoung Lee, Amy N Yates, Daniel Zhou, Timothee Kheyrkhah, Jon Fiscus, et al

  11. [19]

    Yates, Andrew Delgado, Daniel Zhou, Timothee Kheyrkhah, Jeff Smith, and Jonathan Fiscus

    Haiying Guan, Mark Kozak, Eric Robertson, Yooyoung Lee, Amy N. Yates, Andrew Delgado, Daniel Zhou, Timothee Kheyrkhah, Jeff Smith, and Jonathan Fiscus. 2019. MFC Datasets: Large-Scale Benchmark Datasets for Media Forensic Challenge Evaluation. In2019 IEEE Winter Applications o...

    2019

  12. [20]

    Kaiming He, Xiangyu Zhang, Shaoqing Ren, and Jian Sun. 2015. Deep Residual Learning for Image Recognition.CoRR abs/1512.03385 (2015). arXiv:1512.03385 http://arxiv.org/abs/1512.03385

    2015 arXiv

  13. [21]

    https://doi.org/10.6028/NIST.IR.8377

    User guide for nist media forensic challenge (mfc) datasets. https://doi.org/10.6028/NIST.IR.8377

    doi

  14. [22]

    Silvan Heller, Luca Rossetto, and Heiko Schuldt. 2018. The PS-Battles Dataset – an Image Collection for Image Manipulation Detection.CoRRabs/1804.04866 (2018). arXiv:1804.04866 http://arxiv.org/abs/1804.04866

    2018 arXiv

  15. [23]

    Yinqiu Huang, Min Gao, Kai Shu, Chenghua Lin, Jia Wang, and Wei Zhou. 2024. EML: Emotion-Aware Meta Learning for Cross-Event False Information Detection.ACM Trans. Knowl. Discov. Data18, 8 (2024), 185:1–185:25. https: //doi.org/10.1145/3661485

    2024 doi

  16. [24]

    Sifeng He, Yue He, Minlong Lu, Chen Jiang, Xudong Yang, Feng Qian, Xiaobo Zhang, Lei Yang, and Jiandong Zhang

  17. [25]

    Lyndon Kennedy and Shih-Fu Chang. 2008. Internet Image Archaeology: Automatically Tracing the Manipulation History of Photographs on the Web. InProceedings of the 16th ACM International Conference on Multimedia(Vancouver, British Columbia, Canada)(MM ’08). Association for Comp...

    2008

  18. [26]

    Joseph B Kruskal. 1956. On the shortest spanning subtree of a graph and the traveling salesman problem.Proceedings of the American Mathematical society7, 1 (1956), 48–50. https://doi.org/10.1090/S0002-9939-1956-0078686-7

    1956 doi

  19. [27]

    Myung-Joon Kwon, Seung-Hun Nam, In-Jae Yu, Heung-Kyu Lee, and Changick Kim. 2022. Learning JPEG compression artifacts for image manipulation detection and localization.International Journal of Computer Vision130, 8 (2022), 1875–1895. https://doi.org/10.1007/s11263-022-01617-5

    2022 doi

  20. [28]

    Shubham Jain, Ana-Maria Cretu, Antoine Cully, and Yves-Alexandre de Montjoye. 2023. Deep perceptual hashing algorithms with hidden dual purpose: when client-side scanning does facial recognition. In44th IEEE Symposium on Security and Privacy, SP 2023, San Francisco, CA, USA, M...

    2023

  21. [29]

    Xiaoping Liang, Zhenjun Tang, Ziqing Huang, Xianquan Zhang, and Shichao Zhang. 2023. Efficient Hashing Method Using 2D-2D PCA for Image Copy Detection.IEEE Trans. Knowl. Data Eng.35, 4 (2023), 3765–3778. https://doi.org/10. 1109/TKDE.2021.3131188

    2023

  22. [30]

    Shayne Longpre, Robert Mahari, Naana Obeng-Marnu, William Brannon, Tobin South, Katy Ilonka Gero, Alex Pentland, and Jad Kabbara. 2024. Position: Data Authenticity, Consent, & Provenance for AI are all broken: what will it take to fix them?. InForty-first International Confere...

    2024

  23. [31]

    Zhan Luo, Lan Zhang, Jiewei Lai, Xinming Wang, and Chen Tang. 2023. An Efficient Partial Video Copy Detection for a Large-scale Video Database. In9th International Conference on Big Data Computing and Communications, BigCom 2023, Hainan, China, August 4-6, 2023. IEEE, 117–125....

    2023

  24. [32]

    Seongwon Lee, Suhyeon Lee, Hongje Seong, and Euntai Kim. 2023. Revisiting Self-Similarity: Structural Embedding for Image Retrieval. InProceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR). 23412–23421

    2023

  25. [33]

    Rui Ma, Mengxi Guo, Yi Hou, Fan Yang, Yuan Li, Huizhu Jia, and Xiaodong Xie. 2022. Towards Blind Watermarking: Combining Invertible and Non-invertible Mechanisms. InMM ’22: The 30th ACM International Conference on Multimedia, Lisboa, Portugal, October 10 - 14, 2022, João Magal...

    2022

  26. [34]

    Bowyer, Patrick J

    Daniel Moreira, Aparna Bharati, Joel Brogan, Allan Pinto, Michael Parowski, Kevin W. Bowyer, Patrick J. Flynn, Anderson Rocha, and Walter J. Scheirer. 2018. Image Provenance Analysis at Scale.IEEE Transactions on Image Processing27, 12 (2018), 6109–6123. https://doi.org/10.110...

    2018

  27. [35]

    Xin Mu, Ming Pang, and Feida Zhu. 2024. Data Provenance via Differential Auditing.IEEE Trans. Knowl. Data Eng.36, 10 (2024), 5066–5079. https://doi.org/10.1109/TKDE.2023.3334821

    2024

  28. [36]

    Li Ma, Yifan Zhao, Peixi Peng, and Yonghong Tian. 2024. Sensitivity Decouple Learning for Image Compression Artifacts Reduction.IEEE Trans. Image Process.33 (2024), 3620–3633. https://doi.org/10.1109/TIP.2024.3403034

    2024

  29. [37]

    Yakun Niu, Benedetta Tondi, Yao Zhao, Rongrong Ni, and Mauro Barni. 2021. Image Splicing Detection, Localization and Attribution via JPEG Primary Quantization Matrix Estimation and Clustering.IEEE Transactions on Information Forensics and Security16 (2021), 5397–5412. https://...

    2021

  30. [38]

    Oliveira, P

    A. Oliveira, P. Ferrara, A. De Rosa, A. Piva, M. Barni, S. Goldenstein, Z. Dias, and A. Rocha. 2014. Multiple parenting identification in image phylogeny. In2014 IEEE International Conference on Image Processing (ICIP). IEEE, Paris, France, 5347–5351. https://doi.org/10.1109/I...

    2014

  31. [39]

    Zoe Papakipos and Joanna Bitton. 2022. AugLy: Data Augmentations for Robustness. arXiv:2201.06494 [cs.AI] https://arxiv.org/abs/2201.06494

    2022 arXiv

  32. [40]

    Eric Nguyen, Tu Bui, Viswanathan Swaminathan, and John Collomosse. 2021. OSCAR-Net: Object-centric Scene Graph Attention for Image Attribution. In2021 IEEE/CVF International Conference on Computer Vision (ICCV). IEEE, Montreal, QC, Canada, 14479–14488. https://doi.org/10.1109/...

    2021

  33. [41]

    De Rosa, F

    A. De Rosa, F. Uccheddu, A. Costanzo, A. Piva, and M. Barni. 2010. Exploring image dependencies: a new challenge in image forensics. InMedia Forensics and Security II, Nasir D. Memon, Jana Dittmann, Adnan M. Alattar, and Edward J. Delp III (Eds.), Vol. 7541. International Soci...

    2010 doi

  34. [42]

    Leonard Rosenthol, Andy Parsons, Eric Scouten, Jatin Aythora, Bruce Mac-Cormack, Paul England, Marc Levallee, Jonathan Dotan, et al. 2020. Content authenticity initiative (CAI): Setting the standard for content attribution.Technical report, Adobe Inc(2020)

    2020

  35. [43]

    Naoya Sogi, Takashi Shibata, and Makoto Terao. 2024. Object-Aware Query Perturbation for Cross-Modal Image-Text Retrieval. InComputer Vision - ECCV 2024 - 18th European Conference, Milan, Italy, September 29-October 4, 2024, , Vol. 1, No. 1, Article . Publication date: Septemb...

    2024

  36. [44]

    Jayanta Paul, Siddhartha Mallick, Abhijit Mitra, Anuska Roy, and Jaya Sil. 2025. Multi-Modal Twitter Data Analysis for Identifying Offensive Posts Using a Deep Cross Attention based Transformer Framework.ACM Trans. Knowl. Discov. Data(Jan. 2025). https://doi.org/10.1145/371307...

    2025 doi

  37. [45]

    Muhammad Umair, Athman Bouguettaya, Abdallah Lakhdari, Mourad Ouzzani, and Yuyun Liu. 2024. Exif2Vec: A Framework to Ascertain Untrustworthy Crowdsourced Images Using Metadata.ACM Trans. Web18, 3 (2024), 31:1–31:27. https://doi.org/10.1145/3645094

    2024 doi

  38. [46]

    Guanjie Wang, Zehua Ma, Chang Liu, Xi Yang, Han Fang, Weiming Zhang, and Nenghai Yu. 2024. MuST: Robust Image Watermarking for Multi-Source Tracing. InThirty-Eighth AAAI Conference on Artificial Intelligence, AAAI 2024, Thirty- Sixth Conference on Innovative Applications of Ar...

    2024

  39. [47]

    Tianyun Yang, Ziyao Huang, Juan Cao, Lei Li, and Xirong Li. 2022. Deepfake Network Architecture Attribution. In Thirty-Sixth AAAI Conference on Artificial Intelligence, AAAI 2022, Thirty-Fourth Conference on Innovative Applications of Artificial Intelligence, IAAI 2022, The Tw...

    2022 doi

  40. [48]

    NIST MediFor Team. 2017. Nimble Challenge 2017 Evaluation Plan. https://www.nist.gov/system/files/documents/ 2017/09/07/nc2017evaluationplan_20170804.pdf

    2017

  41. [49]

    Jiaxiang You, Yuanman Li, Jiantao Zhou, Zhongyun Hua, Weiwei Sun, and Xia Li. 2021. A Transformer Based Approach for Image Manipulation Chain Detection. InProceedings of the 29th ACM International Conference on Multimedia(Virtual Event, China)(MM ’21). Association for Computin...

    2021

  42. [50]

    Keyang Zhang, Chenqi Kong, Shiqi Wang, Anderson Rocha, and Haoliang Li. 2024. Image Provenance Analysis via Graph Encoding with Vision Transformer.arXiv preprint arXiv:2408.14170(2024). , Vol. 1, No. 1, Article . Publication date: September 2025

    2024 arXiv

  43. [52]

    Shuhei Yokoo. 2021. Contrastive learning with large memory bank and negative embedding subtraction for accurate copy detection.arXiv preprint arXiv:2112.04323(2021). https://arxiv.org/abs/2112.04323

    2021 arXiv

  44. [2013]

    https://doi.org/10.1109/CVPR.2013.379

    IEEE Computer Society, 2946–2953. https://doi.org/10.1109/CVPR.2013.379

    2013 doi

  45. [2021]

    Transformation-Aware Embeddings for Image Provenance.IEEE Transactions on Information Forensics and , Vol. 1, No. 1, Article . Publication date: September 2025. 24 Jiewei Lai, Lan zhang, Chen Tang, and Pengcheng Sun Security16 (2021), 2493–2507. https://doi.org/10.1109/TIFS.20...

    2021

  46. [2023]

    TransVCL: Attention-Enhanced Video Copy Localization Network with Flexible Supervision. InThirty-Seventh AAAI Conference on Artificial Intelligence, AAAI 2023, Thirty-Fifth Conference on Innovative Applications of Artificial Intelligence, IAAI 2023, Thirteenth Symposium on Edu...

    2023 doi

Pith tools

Reviewed August 6, 2026 · model on record in the stance chip above.