- Rating
- 3(+1)
- Confidence
- 3
- Informativeness
- 2
- Presentation
- 2
- Soundness
- 2
Read review
Paper summary
The paper proposes TALA, a surrogate-free query-based attack on image forgery localization. It uses victim feedback to adapt perturbation location, size, and type. Experiments across three detectors and five datasets report higher attack success rates than existing baselines in most evaluated settings.
Critical strengths
Dense localization feedback provides spatial information that TALA uses to guide its search without training a surrogate. The results across multiple detectors and datasets support the value of this approach when target feedback is available.
Critical weaknesses
TALA requires repeated access to the deployed detector’s pixel-level probability maps, which may be difficult to obtain in practice. Transfer-based attacks can prepare adversarial images offline without target feedback. Without evidence on realistic access and generation time, the practical threat posed by TALA is not well established.
Questions for authors
-
In practice, attackers may not know which forgery detector is deployed. TALA nevertheless depends on feedback from the deployed detector to optimize its perturbations. This access assumption limits its applicability, and the setting is not clearly more realistic than generating adversarial images on surrogate models and transferring them to an unknown target.
-
The paper explicitly notes that attack transferability depends strongly on the detector and input image, but does not evaluate the cross-detector transferability of TALA. For example, it remains unclear whether adversarial images generated against TruFor are effective against FOCAL or IFOSN.
-
The Integrated Gradients visualizations lack implementation details, including the output being attributed and the quantity with respect to which gradients are computed. This makes the actual meaning of the heatmaps difficult to assess.
-
Transfer-based attacks can generate adversarial images offline, whereas TALA requires repeated victim queries during attack generation. Its computational complexity and actual generation time are therefore important to assessing its practical threat. The paper does not compare these costs, leaving the time overhead of its online search unclear.
-
Perturbation budgets are selected separately across experimental settings, and TALA is allowed larger perturbations in some cases to improve ASR. The rule for balancing attack success and visual quality is not quantified, making the source of the reported advantage difficult to determine.
-
The method relies on several hand-designed rules and numerous hyperparameters, but the intuition and motivation behind these choices are not sufficiently explained.
13 comments
Response to Reviewer Jd2Y (Part 2/8)
W1 & Q1 & Q4: Generation time (continued).
5. Generation time (Table R3-1). TALA generates adversarial images 3–9× faster per image than the transfer baselines PGN and MEF [1] (TIFS'26; shown to transfer better than PGN and the CVPR'25 method OPS [2] in image classification tasks [1]), while—under exactly the same configurations—achieving comparable ASR and LPIPS in the cross-detector transfer setting (response to Q2, Table R3-4 below) and significantly higher ASR at lower LPIPS in the query setting (Table R3-3).
Protocol. All methods are timed on the same single GPU (NVIDIA RTX 6000 Ada) over identical images, measuring only adversarial-example generation (no I/O or data loading); each call is wrapped with torch.cuda.synchronize() around a wall-clock timer, and the first image is discarded as warm-up. We report the mean and median over the 15 detector–dataset settings. Timed regions and time-critical parameters: for PGN and MEF, the full optimization loop at their officially recommended steps, including all augmented samples per step, as implemented in their official code; for TALA, the full query loop until success or query budget exhaustion at .
Table R3-1. Per-image generation time.
| Method | Access | Mean (s) | Median (s) | vs. TALA |
|---|---|---|---|---|
| TALA (Ours) | query (online) | 12.8 | 8.4 | 1× |
| MEF | transfer (offline) | 39.2 | 22.4 | 3.1× slower |
| PGN | transfer (offline) | 112.4 | 65.4 | 8.8× slower |
The gap has a structural cause: TALA terminates as soon as an image succeeds, whereas transfer attacks always run their full fixed schedule of surrogate forward/backward passes, where is the number of optimization steps and the number of augmented input samples drawn per step (both set to 20 by their official configurations). The medians show the gap is not driven by the large-image IMD2020 outlier that inflates all methods.
Two further costs also favor TALA:
- peak GPU memory is low, as TALA performs no back-propagation and stores no activation graphs;
- transfer attacks additionally carry the one-time offline cost of training/aligning a surrogate, which TALA avoids entirely. The online search thus introduces no time penalty—per image, it is in fact cheaper than offline transfer generation.
[1] Qiu C, Li A, Duan Y, et al. Boosting adversarial transferability with low-cost optimization via maximin expected flatness. IEEE TIFS, 2026.
[2] Guo Y, et al. Boosting adversarial transferability through augmentation in hypothesis space. CVPR 2025, pp. 19175–19185.
Response to Reviewer Jd2Y (Part 3/8)
W1 & Q1 & Q4: Generation computational complexity (continued).
6. Computational complexity (Table R3-2). We characterize cost hardware-independently by the number of network evaluations, reporting forward and backward passes separately. TALA performs no back-propagation: at most victim forward passes per image (average ) and zero backward passes. MEF runs surrogate forward and 400 backward passes; PGN also uses and additionally computes a gradient-norm penalty (an extra gradient per sample), i.e., a comparable-or-larger fixed schedule, consistent with its wall-clock over MEF.
Table R3-2. Per-image network evaluations.
| Method | Victim forwards | Surrogate forwards | Surrogate backwards |
|---|---|---|---|
| TALA (Ours) | (avg ) | 0 | 0 |
| MEF | 0 | 400 | 400 |
| PGN | 0 |
The two families spend entirely different resources: TALA, a small number of victim queries with no gradient computation; transfer attacks, hundreds of surrogate forward and backward passes with no victim access.
Summary. Feedback that is the detector's standard output, no knowledge of the deployed model required, at most 50 queries, 3–9× faster and lower-memory generation, and no surrogate-training cost—TALA constitutes a realistic, practically relevant threat model, complementary to transfer-based attacks.
Table R3-3. Query-setting attack effectiveness and perceptual quality on the large-scale evaluation — images with initial localization F1 under a maximum query budget (up to 1000 per detector–dataset pair). Each entry reports ASR (%) / LPIPS .
| Victim | Method | CASIAv1 | COCOGlide | COVERAGE | IMD2020 | FaceApp |
|---|---|---|---|---|---|---|
| TruFor | PGN | 18.1 / | / 0.1774 | / 0.1949 | / 0.2580 | 94.7 / |
| MEF | / 0.0477 | 45.9 / | 25.5 / | 36.6 / | / 0.1871 | |
| Ours | 61.4 / 0.0350 | 96.4 / 0.0969 | 85.5 / 0.1274 | 74.5 / 0.1905 | 98.1 / 0.1493 | |
| IFOSN | PGN | / | / | / | / | / 0.2546 |
| MEF | 65.3 / 0.1640 | 72.9 / 0.1294 | 63.2 / 0.0946 | 38.8 / 0.1992 | 88.8 / | |
| Ours | 82.4 / 0.1549 | 95.8 / 0.1072 | 100.0 / 0.0789 | 58.9 / 0.1793 | 93.7 / 0.2282 | |
| FOCAL | PGN | 37.1 / | / 0.1678 | / 0.1890 | 23.6 / | 89.1 / |
| MEF | / 0.1366 | 49.8 / | 28.1 / | / 0.1867 | / 0.1334 | |
| Ours | 59.3 / 0.0969 | 89.2 / 0.0956 | 90.2 / 0.1155 | 32.2 / 0.1364 | 97.4 / 0.0763 |
Response to Reviewer Jd2Y (Part 1/8)
We truly appreciate your careful reading and constructive feedback. Your comments are well taken and they have helped us make the practical threat and the method itself much clearer. We address each point in detail below, with new experiments and quantitative evidence, and we hope these responses resolve your concerns.
We would be glad to clarify anything further during the discussion. The revised PDF now incorporates the new experimental results and the corresponding changes in wording discussed below, highlighted in blue for ease of checking.
W1 & Q1 & Q4: Threat-model positioning and realistic access.
1. Positioning: two standard, complementary threat models.
2. The access assumption is weaker than it may appear. TALA observes only the victim’s standard localization output (predicted map or confidence score): no weights, no gradients, no internal state, and no knowledge of which detector is deployed. Not knowing the deployed model (Q1) therefore poses no obstacle—TALA adapts to whatever responds, which is exactly the setting it targets. Transfer attacks, generated offline, have no such mechanism when a perturbation fails on the deployed model.
3. Such access exists in real deployments. Baidu's image-tampering-detection API returns a tampering heatmap (
return_heatmap=true) and an image-level confidence score (detect_proportion=true) [Baidu API]; Copyleaks' AI-image-detection API returns spatial detection masks [Copyleaks API]. Spatially-resolved detection feedback is therefore not hypothetical.4. Research value beyond immediate deployability. As in classification, output-only query attacks chart the security boundary of detectors when only their outputs are exposed, and form a natural progression toward harder hard-label settings. We hope this axis proves worth exploring in its own right, beyond the specifics of any one deployment scenario.