Pith. sign in

REVIEW 5 major objections 7 minor 1 cited by

Exploring the Robustness of AI-Driven Tools in Digital Forensics: A Preliminary Study

T0 review · 5 major / 7 minor · reviewed 2026-08-12 · deepseek-v4-flash

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

Pith's one-line read A preliminary test of Magnet AI and Excire Photo AI shows they misclassify sexual images, miss chat topics, and match some deepfakes to the real person, so analysts cannot trust automated labels.

desk verdict Useful first look at two commercial forensic AI tools, but the robustness claim is weakened because the tools are tested outside their intended scope and every disagreement with manual labels is treated as a misclassification. read the letter →

arxiv 2412.01363 v1 pith:6FMTY5E7 submitted 2024-12-02 cs.CV

Silvia Lucia Sanna , Leonardo Regano , Davide Maiorca , Giorgio Giacinto This is my paper · ORCID
classification cs.CV
keywords digitalforensicsAIrobustnessblack-boxevaluationadversarialexamplesnuditydetectiondeepfakefacerecognitionanti-forensics
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

This paper attempts to show that the artificial-intelligence components inside two commercial digital forensics products—Magnet AI in Magnet Axiom and Excire Photo AI in X-Ways Forensics—are not robust enough to label evidence reliably. Using a self-built dataset of roughly 200 nudity images, 200 face images, and three test chats, the authors report that some sexual images are not flagged as nudity, some harmless objects such as a grape are flagged as drugs, and some deepfakes are matched to the real actor. The paper concludes that these tools produce both false negatives and false positives on non-standard inputs and that forensic analysts must still check every automated label by hand. This matters because missed or invented labels in a criminal investigation can hide evidence, waste analyst time, or skew a case.

What carries the argument

The argument is carried by a black-box evaluation of two proprietary classifiers. Magnet AI is the module inside Magnet Axiom that labels pictures and chat messages with categories such as nudity, drugs, and weapons; Excire Photo AI is the module inside X-Ways Forensics that detects objects and matches faces to known people. The authors built a balanced homemade dataset—200 nudity images spanning real porn, teenage-looking actors, manga, and painted dresses; 200 face images of actors, their children, and people before and after cosmetic or gender-affirming surgery; and 3 chat conversations about drugs, weapons, and sex—loaded it into the tools, and compared each tool's output against human-expected labels, treating a perfect classifier as one that achieves 100% detection in every category. This comparison yields the false-positive and false-negative counts that support the robustness claim.

What would settle it

Obtain the tools' confidence scores and documented category definitions, or adjust their detection thresholds: if the supposedly missed nudity images and the supposedly wrong drug images change labels when the threshold is lowered, or fall outside the vendors' stated category scope, then the reported 'misclassifications' are not evidence of fragility but artifacts of the test's assumed ground truth.

Watch

Extended reading notes

Core claim

The central claim, stated in the authors' own discussion, is that 'these tools are not robust enough in the detection as shown by the high number of misclassifications.' Concretely: Magnet AI fails to mark some sexual images (including some real pornography frames and manga) as nudity, while also flagging clearly innocent pictures, such as a grape, as drugs; Excire Photo AI recognizes some deepfakes as the real person, produces many false positives when queried by actor name or by face similarity, and sometimes matches unrelated people to a queried face. The authors answer all three of their research questions affirmatively: the tools have nontrivial false-positive rates, deepfake faces are sometimes classified as the original person, and these failures imply that adversarial or non-standard content could be used for anti-forensics purposes.

Load-bearing premise

The load-bearing premise is that the authors' manual labels are the correct ground truth and that a perfect classifier should achieve 100% detection in every category; if the vendors' category definitions or confidence thresholds differ from that standard, the reported misclassifications could be artifacts of the evaluation setup, and the paper itself acknowledges the dataset is small and self-built.

Editorial extensions

If this is right

  • If these results are typical, forensic analysts cannot rely on Magnet AI or Excire Photo AI labels without manually reviewing every flagged and unflagged item.
  • The specific failure patterns suggest the classifiers rely on surface cues rather than robust semantic understanding, since painted bodies are counted as nudity, a grape as a drug, and unrelated faces as the queried person.
  • The failures imply an anti-forensics avenue: a suspect could hide illicit material by using non-standard but human-recognizable depictions, or by using deepfakes that the face matcher accepts as the real person.
  • The paper's recommendations—adding explainability, building an adversarial testing platform, and retraining on broader cases—become concrete requirements if these error rates hold.
  • The chat-analysis feature's failure to flag any explicit or metaphorical message suggests natural-language understanding in current forensic tools is too weak to assist topic detection in investigations.

Reading between the lines

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

  • The same black-box evaluation method could be standardized into a benchmark for any commercial forensic AI; a public test set of non-standard images would likely reveal similar fragility in other vendors' classifiers.
  • The deepfake face-matching results point beyond forensics: face recognition systems trained on celebrity images may be generally vulnerable to synthetic faces, with implications for identity fraud and surveillance.
  • A testable extension would be to vary image compression, filters, and small adversarial perturbations on the same datasets and measure how detection rates swing; large swings would indicate the robustness problem is even more severe than this coarse test shows.
  • The 'grape as drugs' false positive hints that the drug classifier keys on color and texture; identifying those features could allow crafting targeted evasion images, which the authors do not attempt here.
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 / 7 minor

Summary. The paper presents a preliminary black-box evaluation of two commercial digital-forensics tools with AI components: Magnet AI (integrated in Magnet Axiom) and Excire Photo AI (used by X-Ways Forensics). The authors created a custom dataset of about 200 nudity-related images (real porn, teenager-appearance actors, manga, painted dresses), 200 face-recognition images (actors, deepfakes, children, post-surgery people), and three real Facebook Messenger chats about drugs, weapons, and sex, which were acquired with the Magnet toolchain. They report detection percentages and false-positive counts per category, and conclude that the tools are not robust enough and that forensic analysts must manually verify AI outputs. The paper frames this as a preliminary study and lists future work including larger datasets, LLM-generated chat data, and explainable AI.

Significance. The topic is practically important: these commercial tools are deployed in real forensic investigations, and a black-box evaluation that documents concrete failures (e.g., a grape flagged as a drug, deepfakes of Brad Pitt returned in a face query) can inform both practitioners and tool vendors. The authors also took steps that strengthen the study: they used a real Android memory acquisition for the chat experiments, performed multiple acquisition runs to check stability, and explicitly considered ethical constraints by not publishing child-abuse-related content. However, the quantitative claims are currently not fully supported by the evidence: the dataset is small and hand-selected, the ground truth is the authors' own semantic labels rather than an independently validated standard, and the paper does not establish that the tested inputs fall within the tools' intended categories. The work would be an appropriate and useful empirical case study if the conclusions were reframed accordingly and the missing validation steps were provided.

major comments (5)
  1. [Section 5.1, Figure 1] The statement that 'a perfect classifier should have a 100% detection rate for each category' is an untested premise. The images are hand-selected by the authors, and no independent ground-truth validation (e.g., a second labeling process or a comparison with a reference classifier on a control set) is provided. As a result, the percentages in Figures 1-3 conflate 'deviation from the authors' expectations' with 'misclassification.' A concrete test would be to run an additional set of clearly in-domain images (ordinary photographic nudity without painted or drawn content) to establish a baseline, or to measure inter-annotator agreement on the category labels.
  2. [Section 6.2, Figures 6-7] For Excire Photo AI, the paper treats any deepfake face returned by a name query as a false positive, and in Section 6.3 (RQ2) concludes that 'some deepfakes are recognized as the real person.' But Excire Photo AI is a face-matching/similarity-search tool, not a deepfake detector; returning faces visually similar to the query is its stated function. The failure claim is only meaningful if Excire explicitly claims to distinguish genuine from synthetic faces, or if the returned deepfake does not actually resemble the query. I recommend either consulting the tool's documentation for its intended scope or adding a control condition with non-deepfake lookalikes to show that the tool is specifically confused by the synthetic nature of the images.
  3. [Section 6.1, Figure 4] The nudity experiments include 'dresses with painted nudes' and manga images. The authors infer from NudeNet's better performance that Magnet AI may rely on skin-exposure heuristics, but they do not establish that Magnet AI's 'nudity' category is intended to cover non-photographic or painted nudity. Without this scope check, the undetected manga and painted-dress images in Figure 1 cannot be attributed to non-robustness; they may simply be out-of-scope inputs. The paper should report what the vendors' documentation or public materials say about the covered categories, and ideally rerun the analysis on a clearly in-scope control set.
  4. [Section 5, Figures 1-3] The reported percentages lack error bars, confidence intervals, or significance tests. With only 50 images per category, a few images correspond to a 6-8% swing, and the chat experiment is further contaminated by 'other default pictures found in the Android device memory dump' (Section 6.1) that were not part of the controlled inputs. The load-bearing claim of a 'high number of misclassifications' (Section 6.3, RQ1) would be considerably strengthened by reporting per-image raw counts, selection criteria, and a breakdown by image type, rather than aggregate bars.
  5. [Section 5.1, chat analysis] The paper states that the 'Analyze chats with AI' feature 'was not able to flag any of our chat messages' but does not show the actual chat contents, the tool's confidence scores, or a table of which messages were sent. Since the authors intentionally used slang and metaphors, it is impossible for the reader to determine whether the messages were within the tool's detection scope. A supplementary table listing representative messages and the tool's outputs (or lack thereof) is needed to make this finding verifiable.
minor comments (7)
  1. [Abstract] The phrase 'marks it at as suspicious' is grammatically incorrect; it should be 'marks it as suspicious.' Also, 'other 100 sent in 3 chats' should specify what the 100 refers to (messages and/or images) and how they are split across the three chats.
  2. [Section 2 heading] The heading 'Forensic T orensics Tools' contains a typo: it should be 'Forensics Tools' or 'Digital Forensics Tools'.
  3. [Section 6.1] The sentence 'The undetected pictures are not so different from the detected ones in terms of content and dimension/light/contrast' is vague; please provide quantitative measures (e.g., resolution, aspect ratio) or show the specific images so the claim can be assessed.
  4. [Section 4.1] The term 'interracial' is dated and may carry unintended connotations; consider using 'images with diverse skin tones' to match the stated intention of avoiding bias in the dataset.
  5. [Section 7] In the conclusion, 'full checking their results' should be 'fully checking their results,' and 'human experts must still supervision such algorithms' should be 'must still supervise such algorithms.'
  6. [Figures 2 and 3 captions] The figures label returned non-target images as 'false positives.' In a forensic tool, a returned image that does not match the query is an incorrect result, but the term 'false positive' can be misleading when the tool is a similarity search rather than a binary classifier; please define how the term is used here.
  7. [References] Some references have inconsistent formatting, e.g., Lopes et al., 2009 and the Nowroozi thesis entry; the page ranges and venue names should be checked for consistency with the journal's style.

Circularity Check

0 steps flagged · score 1.0 of 10

No circularity: the robustness conclusion is an empirical black-box finding, not an artifact of self-referential derivation.

full rationale

The paper is an empirical black-box evaluation. Its central claim ('these tools are not robust enough', Section 6.3) is supported by directly observed outputs from Magnet AI and Excire Photo AI on a self-constructed dataset; it is not derived from the tools' own labels, fitted parameters, or a uniqueness argument. The 'perfect classifier should have a 100% detection rate' statement (Section 5.1) is an evaluation benchmark, not an equation into which the results are substituted; it states the authors' criterion, and the observed deviations are contingent measurements. The deepfake face-recognition test reports that some deepfakes are matched to the real person (Section 6.2), which is an observed output, even though the authors themselves expected this because deepfake detection is an open problem. The only self-citation is the background citation of Biggio et al. (2013) for the existence of adversarial examples; it is motivational and does not carry the robustness conclusion. No fitted input is renamed as a prediction, no prior-work uniqueness theorem is imported, and no known result is renamed as unification. Concerns about whether the tools' intended categories match the authors' manual labels are validity and bias issues, not circular reasoning.

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

The central claim rests on the authors' subjective labeling and an assumed perfect-classifier standard. There are no fitted numerical parameters, but the evaluation framework contains these unverified assumptions.

assumptions (3)
  • domain assumption The authors' manual image labels are the correct ground truth for evaluating the tools.
    Section 4.1 describes creating the dataset and the paper judges tool outputs against human expectations (e.g., the grape is called a false positive). No independent ground truth or benchmark is used.
  • ad hoc to paper A perfect forensic classifier should detect 100% of the hand-selected images in each category.
    Section 5.1 states 'a perfect classifier should have a 100% detection rate for each category.' This is an ad hoc yardstick that is not derived from the tools' specifications or any standard.
  • domain assumption The tools are expected to recognize drawn, painted, or unnatural nudity and to match deepfakes to real identities.
    The paper assumes these are failures of the tools. The tools' public documentation lists categories but does not promise coverage of non-photorealistic or manipulated content.

how reviews work

0 comments
Cite this review

Pith. "Pith review of Exploring the Robustness of AI-Driven Tools in Digital Forensics: A Preliminary Study." pith.science (2026). https://pith.science/paper/6FMTY5E7

@misc{pith2026241201363,
  author       = {Pith},
  title        = {Pith review of: Exploring the Robustness of AI-Driven Tools in Digital Forensics: A Preliminary Study},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/6FMTY5E7}},
  note         = {Machine review of arXiv:2412.01363}
}
abstract

Nowadays, many tools are used to facilitate forensic tasks about data extraction and data analysis. In particular, some tools leverage Artificial Intelligence (AI) to automatically label examined data into specific categories (\ie, drugs, weapons, nudity). However, this raises a serious concern about the robustness of the employed AI algorithms against adversarial attacks. Indeed, some people may need to hide specific data to AI-based digital forensics tools, thus manipulating the content so that the AI system does not recognize the offensive/prohibited content and marks it at as suspicious to the analyst. This could be seen as an anti-forensics attack scenario. For this reason, we analyzed two of the most important forensics tools employing AI for data classification: Magnet AI, used by Magnet Axiom, and Excire Photo AI, used by X-Ways Forensics. We made preliminary tests using about $200$ images, other $100$ sent in $3$ chats about pornography and teenage nudity, drugs and weapons to understand how the tools label them. Moreover, we loaded some deepfake images (images generated by AI forging real ones) of some actors to understand if they would be classified in the same category as the original images. From our preliminary study, we saw that the AI algorithm is not robust enough, as we expected since these topics are still open research problems. For example, some sexual images were not categorized as nudity, and some deepfakes were categorized as the same real person, while the human eye can see the clear nudity image or catch the difference between the deepfakes. Building on these results and other state-of-the-art works, we provide some suggestions for improving how digital forensics analysis tool leverage AI and their robustness against adversarial attacks or different scenarios than the trained one.

Figures

Figures reproduced from arXiv: 2412.01363 by the authors.

Figure 2
Figure 2. The bar chart shows the percentage accuracy for the chat [PITH_FULL_IMAGE:figures/full_fig_p007_2.png] view at source ↗
Figure 1
Figure 1. The bar chart shows the percentage of nude images de [PITH_FULL_IMAGE:figures/full_fig_p007_1.png] view at source ↗
Figure 3
Figure 3. The bar chart shows the accuracy of the Excire Photo AI [PITH_FULL_IMAGE:figures/full_fig_p007_3.png] view at source ↗
Figures from the paper (5 more)
Figure 5
Figure 5. Figure 5: The image shows the pictures detected as drugs but that [PITH_FULL_IMAGE:figures/full_fig_p008_5.png]
Figure 6
Figure 6. Figure 6: Obtained results for querying Brad Pitt in the Excire Photo [PITH_FULL_IMAGE:figures/full_fig_p008_6.png]
Figure 8
Figure 8. Figure 8: Overall results about parents’ surname querying in the [PITH_FULL_IMAGE:figures/full_fig_p009_8.png]
Figure 9
Figure 9. Figure 9: Results by selecting Shiloh’s face and looking for face sim [PITH_FULL_IMAGE:figures/full_fig_p009_9.png]
Figure 6
Figure 6. Figure 6: We expected this result since deepfake detection [PITH_FULL_IMAGE:figures/full_fig_p010_6.png]

Discussion (0). Continue with ORCID to comment.

Forward citations

Cited by 1 Pith paper

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

  1. MoLIFE: Methodology, Technologies, and Challenges for Mobile Live Intelligent Forensics Examination

    cs.CR 2026-07 conditional novelty 5.0 of 10

    A mobile Digital Twin with super-user privileges can replicate a real device's forensic data content at 95-100% similarity, enabling privileged forensic analysis without modifying the physical evidence.

Reference graph

Works this paper leans on

43 extracted references · 21 canonical work pages · cited by 1 Pith paper

  1. [1]

    , author Grzonkowski, S

    author Aditya, K. , author Grzonkowski, S. , author Lekhac, N.A. , year 2018 . title Enabling trust in deep learning models: A digital forensics case study . journal 2018 17th IEEE International Conference On Trust, Security And Privacy In Computing And Communications/ 12th IEEE International Conference On Big Data Science And Engineering (TrustCom/BigDat...

    arXiv 2018

  2. [2]

    , author Nawaz, M

    author Adnan, A. , author Nawaz, M. , year 2016 . title Rgb and hue color in pornography detection . journal Springer , pages 1041--1050 :10.1007/978-3-319-32467-8_90

    work page doi:10.1007/978-3-319-32467-8_90 2016

  3. [3]

    , author Abdul Karim, H

    author AlDahoul, N. , author Abdul Karim, H. , author Lye Abdullah, M.H. , author Ahmad Fauzi, M.F. , author Ba Wazir, A.S. , author Mansor, S. , author See, J. , year 2021 . title Transfer detection of yolo to focus cnn’s attention on nude regions for adult content detection . journal Symmetry volume 13 . :10.3390/sym13010026

    work page doi:10.3390/sym13010026 2021

  4. [4]

    , author Elgibreen, H

    author Almutairi, Z. , author Elgibreen, H. , year 2022 . title A review of modern audio deepfake detection methods: Challenges and future directions . journal Algorithms volume 15 . :10.3390/a15050155

    doi:10.3390/a15050155 2022

  5. [5]

    , author Galteri, L

    author Amerini, I. , author Galteri, L. , author Caldelli, R. , author Del Bimbo, A. , year 2019 . title Deepfake video detection through optical flow based cnn . journal Proceedings - 2019 International Conference on Computer Vision Workshop, ICCVW 2019 , pages 1205 – 1207 :10.1109/ICCVW.2019.00152

    arXiv 2019

  6. [6]

    , author Behzadan, V

    author Baggili, I. , author Behzadan, V. , year 2019 . title Founding the domain of ai forensics . journal arXiv http://arxiv.org/abs/1912.06497 arXiv:1912.06497

    arXiv 2019

  7. [7]

    , author Corona, I

    author Biggio, B. , author Corona, I. , author Maiorca, D. , author Nelson, B. , author Šrndić, N. , author Laskov, P. , author Giacinto, G. , author Roli, F. , year 2013 . title Evasion attacks against machine learning at test time . journal Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes ...

    doi:10.1007/978-3-642-40994-3_25 2013

  8. [8]

    , author Rosso, P

    author Bogdanova, D. , author Rosso, P. , author Solorio, T. , year 2012 . title On the impact of sentiment and emotion based features in detecting online sexual predators . journal Proceedings of the Annual Meeting of the Association for Computational Linguistics , pages 110 – 118

    work page 2012

Show all 43 references
  1. [9]

    , author Wagner, D

    author Carlini, N. , author Wagner, D. , year 2017 . title Towards evaluating the robustness of neural networks , p. pages 39 – 57 . :10.1109/SP.2017.49

    2017 doi

  2. [10]

    , author da Silva Eleuterio, P.M

    author de Castro Polastro, M. , author da Silva Eleuterio, P.M. , year 2010 . title Nudetective: A forensic tool to help combat child pornography through automatic nudity detection , pages 349--353 :10.1109/DEXA.2010.74

    2010 doi

  3. [11]

    , author La Cava, S.M

    author Concas, S. , author La Cava, S.M. , author Orrù, G. , author Cuccu, C. , author Gao, J. , author Feng, X. , author Marcialis, G.L. , author Roli, F. , year 2022 . title Analysis of score-level fusion rules for deepfake detection . journal Applied Sciences volume 12 . ht...

    2022 doi

  4. [12]

    , author Ghanem, M

    author Dunsin, D. , author Ghanem, M. , author Ouazzane, K. , author Vassilev, V. , year 2024 . title A comprehensive analysis of the role of artificial intelligence and machine learning in modern digital forensics and incident response . journal Forensic Science International...

    2024

  5. [13]

    , author Suen, C.Y

    author Ebrahimi, M. , author Suen, C.Y. , author Ormandjieva, O. , author Krzyzak, A. , year 2016 . title Recognizing predatory chat documents using semi-supervised anomaly detection . journal IS and T International Symposium on Electronic Imaging Science and Technology :10.23...

    2016

  6. [14]

    , author Micheletto, M

    author Gao, J. , author Micheletto, M. , author Orrù, G. , author Concas, S. , author Feng, X. , author Marcialis, G.L. , author Roli, F. , year 2024 . title Texture and artifact decomposition for improving generalization in deep-learning-based deepfake detection . journal Eng...

    2024

  7. [15]

    , author Revano, T.F

    author Garcia, M.B. , author Revano, T.F. , author Habal, B.G.M. , author Contreras, J.O. , author Enriquez, J.B.R. , year 2018 . title A pornographic image and video filtering application using optimized nudity recognition and detection algorithm , pages 1--5 :10.1109/HNICEM....

    2018

  8. [16]

    , author Giudice, O

    author Guarnera, L. , author Giudice, O. , author Guarnera, F. , author Ortis, A. , author Puglisi, G. , author Paratore, A. , author Bui, L.M.Q. , author Fontani, M. , author Coccomini, D.A. , author Caldelli, R. , author Falchi, F. , author Gennaro, C. , author Messina, N. ,...

    2022

  9. [17]

    , author Delp, E.J

    author Guera, D. , author Delp, E.J. , year 2018 . title Deepfake video detection using recurrent neural networks . journal Proceedings of AVSS 2018 - 2018 15th IEEE International Conference on Advanced Video and Signal-Based Surveillance :10.1109/AVSS.2018.8639163

    2018

  10. [18]

    , year 2019

    author Gunning, D. , year 2019 . title Darpa's explainable artificial intelligence (xai) program , in: booktitle Proceedings of the 24th International Conference on Intelligent User Interfaces , publisher Association for Computing Machinery , address New York, NY, USA . p. pag...

    2019

  11. [19]

    , author Javed, A.R.R

    author Hamza, A. , author Javed, A.R.R. , author Iqbal, F. , author Kryvinska, N. , author Almadhor, A.S. , author Jalil, Z. , author Borghol, R. , year 2022 . title Deepfake audio detection via mfcc features using machine learning . journal IEEE Access volume 10 , pages 13401...

    2022

  12. [20]

    , author Uddin, M.A

    author Islam, M.M. , author Uddin, M.A. , author Islam, L. , author Akter, A. , author Sharmin, S. , author Acharjee, U.K. , year 2020 . title Cyberbullying detection on social networks using machine learning approaches . journal 2020 IEEE Asia-Pacific Conference on Computer S...

    2020

  13. [21]

    Macedo, F.C

    author J. Macedo, F.C. , author dos Santos, J.A. , year 2018 . title A benchmark methodology for child pornography detection , in: booktitle Graphics, Patterns and Images (SIBGRAPI), 2018 31th SIBGRAPI Conference on , organization IEEE

    2018

  14. [22]

    , author Chang, M.C

    author Li, Y. , author Chang, M.C. , author Lyu, S. , year 2018 . title In ictu oculi: Exposing ai created fake videos by detecting eye blinking . journal 10th IEEE International Workshop on Information Forensics and Security, WIFS 2018 :10.1109/WIFS.2018.8630787

    2018

  15. [23]

    , author De Avila, S.E.F

    author Lopes, A.P.B. , author De Avila, S.E.F. , author Peixoto, A.N.A. , author Oliveira, R.S. , author Coelho, M.D.M. , author Araújo, A.D.A. , year 2009 . title Nude detection in video using bag-of-visual-features . journal Proceedings of SIBGRAPI 2009 - 22nd Brazilian Symp...

    2009 doi

  16. [24]

    , author Bhattacharya, U

    author Mittal, T. , author Bhattacharya, U. , author Chandra, R. , author Bera, A. , author Manocha, D. , year 2020 . title Emotions don't lie: An audio-visual deepfake detection method using affective cues . journal MM 2020 - Proceedings of the 28th ACM International Conferen...

    2020

  17. [25]

    , author Zhang, M

    author Mohanty, M. , author Zhang, M. , author Russello, G. , year 2019 . title A photo forensics-based prototype to combat revenge porn , pages 5--8 :10.1109/MIPR.2019.00009

    2019

  18. [26]

    , author Avila, S

    author Moreira, D. , author Avila, S. , author Perez, M. , author Moraes, D. , author Testoni, V. , author Valle, E. , author Goldenstein, S. , author Rocha, A. , year 2016 . title Pornography classification: The hidden clues in video space–time . journal Forensic Science Inte...

    2016 doi

  19. [27]

    , author Fechine, J.M

    author Moreira, D.C. , author Fechine, J.M. , year 2018 . title A machine learning-based forensic discriminator of pornographic and bikini images , pages 1--8 :10.1109/IJCNN.2018.8489100

    2018

  20. [28]

    , author Eloff, J

    author Ngejane, C. , author Eloff, J. , author Sefara, T. , author Marivate, V. , year 2021 . title Digital forensics supported by machine learning for the detection of online sexual predatory chats . journal Forensic Science International: Digital Investigation volume 36 . :1...

    2021

  21. [29]

    , year 2020

    author Nowroozi, E. , year 2020 . title Machine learning techniques for image forensics in adversarial setting . journal University of Siena :10.13140/RG.2.2.24757.83686

    2020

  22. [30]

    , author La Cava, S.M

    author Panzino, A. , author La Cava, S.M. , author Orr\`u, G. , author Marcialis, G.L. , year 2024 . title Evaluating the integration of morph attack detection in automated face recognition systems , in: booktitle Proceedings of the IEEE/CVF Conference on Computer Vision and P...

    2024

  23. [31]

    , author McDaniel, P

    author Papernot, N. , author McDaniel, P. , author Goodfellow, I. , author Jha, S. , author Celik, Z.B. , author Swami, A. , year 2017 . title Practical black-box attacks against machine learning , p. pages 506 – 519 . :10.1145/3052973.3053009

    2017

  24. [32]

    , author Tan, W

    author Qi, D. , author Tan, W. , author Liu, Z. , author Yao, Q. , author Liu, J. , year 2020 . title A dataset and system for real-time gun detection in surveillance video using deep learning

    2020

  25. [33]

    , author Divvala, S

    author Redmon, J. , author Divvala, S. , author Girshick, R. , author Farhadi, A. , year 2016 . title You only look once: Unified, real-time object detection , in: booktitle 2016 IEEE Conference on Computer Vision and Pattern Recognition (CVPR) , pp. pages 779--788 . :10.1109/...

    2016 doi

  26. [34]

    why should i trust you?

    author Ribeiro, M.T. , author Singh, S. , author Guestrin, C. , year 2016 . title "why should i trust you?": Explaining the predictions of any classifier , in: booktitle Proceedings of the 22nd ACM SIGKDD International Conference on Knowledge Discovery and Data Mining , publis...

    2016

  27. [35]

    , author Grajeda, C

    author Sanchez, L. , author Grajeda, C. , author Baggili, I. , author Hall, C. , year 2019 . title A practitioner survey exploring the value of forensic tools, ai, filtering, & safer presentation for investigating child sexual abuse material (csam) . journal Digital Investigat...

    2019 doi

  28. [36]

    , author dos Santos, E.M

    author Santos, C. , author dos Santos, E.M. , author Souto, E. , year 2012 . title Nudity detection based on image zoning , pages 1098--1103 :10.1109/ISSPA.2012.6310454

    2012

  29. [37]

    , author Breitinger, F

    author Schneider, J. , author Breitinger, F. , year 2023 . title Towards ai forensics: Did the artificial intelligence system do it? journal Journal of Information Security and Applications volume 76 , pages 103517 . :https://doi.org/10.1016/j.jisa.2023.103517

    2023

  30. [38]

    , year 2022

    author Solanke, A.A. , year 2022 . title Explainable digital forensics ai: Towards mitigating distrust in ai-based digital forensics analysis using interpretable models . journal Forensic Science International: Digital Investigation volume 42 , pages 301403 . :https://doi.org/...

    2022

  31. [39]

    , author Razi, A

    author Tariq, M.U. , author Razi, A. , author Badillo-Urquiola, K. , author Wisniewski, P. , year 2019 . title A review of the gaps and opportunities of nudity and skin detection algorithmic research for the purpose of combating adolescent sexting behaviors . journal Springer ...

    2019 doi

  32. [40]

    , author Vera-Rodriguez, R

    author Tolosana, R. , author Vera-Rodriguez, R. , author Fierrez, J. , author Morales, A. , author Ortega-Garcia, J. , year 2020 . title Deepfakes and beyond: A survey of face manipulation and fake detection . journal Information Fusion volume 64 , pages 131 – 148 . :10.1016/j...

    2020 doi

  33. [41]

    , author Li, Y

    author Yang, X. , author Li, Y. , author Lyu, S. , year 2019 . title Exposing deep fakes using inconsistent head poses . journal ICASSP, IEEE International Conference on Acoustics, Speech and Signal Processing - Proceedings volume 2019-May , pages 8261 – 8265 . :10.1109/ICASSP...

    2019

  34. [42]

    , author Fu, R

    author Yi, J. , author Fu, R. , author Tao, J. , author Nie, S. , author Ma, H. , author Wang, C. , author Wang, T. , author Tian, Z. , author Bai, Y. , author Fan, C. , author Liang, S. , author Wang, S. , author Zhang, S. , author Yan, X. , author Xu, L. , author Wen, Z. , a...

    2022

  35. [43]

    , author Xin, Y

    author Yuezun, L. , author Xin, Y. , author Pu, S. , author Honggang, Q. , author Siwei, L. , year 2020 . title Celeb-df: A large-scale challenging dataset for deepfake forensics , in: booktitle IEEE Conference on Computer Vision and Patten Recognition (CVPR)

    2020

Pith tools

Reviewed August 12, 2026 · model on record in the stance chip above.