- Rating
- 3
- Confidence
- 2
- Informativeness
- 2
- Presentation
- 3
- Soundness
- 2
Read review
Paper summary
This paper studies defense against physically realizable, full-surface adversarial camouflage attacks on monocular depth estimation. The authors propose CamouPurifier, a zero-shot preprocessing framework consisting of camouflage localization and texture reconstruction. It first detects candidate vehicle regions and combines DINO-based semantic-aware localization with several handcrafted texture-anomaly cues to identify suspicious camouflage pixels. The localized regions are then removed and optionally reconstructed using Stable Diffusion inpainting guided by Canny edges through ControlNet, followed by Gaussian boundary blending. Experiments in CARLA cover four attacks and six monocular depth estimation models, with additional evaluations across weather conditions, vehicle types, an adaptive attack, and a small-scale physical case study. The reported results show substantial reductions in vehicle-region depth discrepancy compared with no defense and the Jedi, PAD, and APEX baselines. The localization-only and reconstruction-based variants provide different trade-offs between attack suppression, image fidelity, and computational cost.
Critical strengths
-
Clear and easy-to-follow presentation. The paper is generally well organized, and the motivation, threat model, and overall defense pipeline are presented clearly. The decomposition into camouflage localization and texture reconstruction makes the proposed approach intuitive and easy to understand.
-
Meaningful and practically relevant problem setting. Full-surface adversarial camouflage poses a realistic physical threat to monocular depth estimation systems, particularly in autonomous-driving scenarios. Compared with conventional imperceptible perturbations or localized patches, this threat remains relatively underexplored from the defense perspective.
-
Strong empirical performance. CamouPurifier achieves substantial reductions in vehicle-region depth discrepancy across multiple attacks and MDE architectures. The results are generally consistent across different weather conditions and vehicle types, and the method compares favorably with the evaluated Jedi, PAD, and APEX baselines.
-
Broad experimental coverage. The evaluation includes four attack variants, six MDE models, diverse environmental conditions, an adaptive attack, and a physical case study. The ablation study also helps clarify the contributions of the semantic localization, texture anomaly localization, and reconstruction components.
Critical weaknesses
-
Limited practical efficiency. Although the paper motivates the defense in the context of autonomous driving, its runtime is far from real-time. Even the localization-only variant runs at only 2.705 FPS on an A100 GPU, while the complete reconstruction pipeline runs at 0.098 FPS, corresponding to approximately 10 seconds per image. It is also unclear whether these measurements include the downstream MDE inference. Therefore, the claim that the method is suitable for latency-sensitive scenarios is not sufficiently supported.
-
Potential limitations of the evaluation protocol. For each attack–model pair, the authors rank 1,043 samples according to their undefended vehicle-region error and retain only the strongest 500 samples. Since this selection is performed before applying any defense, it does not directly bias the relative comparison among defense methods. However, the resulting evaluation measures recovery conditional on highly successful attacks and does not characterize performance over the original data distribution. Reporting results over all samples, together with attack success rates and per-condition statistics, would provide a more complete evaluation.
-
Several claims are not fully supported by the current evaluation. First, although the paper reports clean-image depth discrepancy, it does not report the localization false-positive rate or predicted mask area on benign vehicles. Because both localization branches use percentile-based thresholds, it would be useful to evaluate whether naturally complex vehicle appearances, such as decals, reflections, or patterned paint, trigger unnecessary purification. Second, the adaptive attack is only briefly described, and the DINO percentile threshold is changed specifically for this experiment rather than fixed beforehand. More details about the adaptive objective and optimization procedure are needed. Finally, while the aggregate results are strong, the defense can degrade performance for some individual attack–model pairs. For example, under MIMDepth and 3D²Fool, the vehicle-region discrepancy increases from 2.860 without defense to 3.771 and 3.209 for the two variants. A clearer discussion of such failure cases would help qualify the overall conclusions.
Questions for authors
Please see the questions raised in the Critical Weaknesses section.
6 comments
Author Response to Reviewer iCcG (Part 2/3)
2. Response to W2
W2: Potential limitations of the evaluation protocol. For each attack–model pair, the authors rank 1,043 samples according to their undefended vehicle-region error and retain only the strongest 500 samples. Since this selection is performed before applying any defense, it does not directly bias the relative comparison among defense methods. However, the resulting evaluation measures recovery conditional on highly successful attacks and does not characterize performance over the original data distribution. Reporting results over all samples, together with attack success rates and per-condition statistics, would provide a more complete evaluation.
We sincerely thank the reviewer for this constructive suggestion. We agree that evaluating only the 500 samples with the largest undefended vehicle-region discrepancies primarily characterizes defense effectiveness against highly successful attacks, rather than performance over the complete data distribution.
Motivation for the original selection. Our original Top-500 evaluation was intended to focus on challenging adversarial scenarios in which camouflage attacks induce substantial depth prediction deviations. As the reviewer correctly noted, the selection was performed before applying any defense and therefore did not directly favor a particular defense method.
Additional evaluation on all scenes. To address this concern, we have extended the evaluation to all 1,043 scenes for Monodepth2 across four adversarial camouflage attacks: 3D2Fool, 3D-APA, 3D-SSAP, and RobCamou. All baseline defenses are evaluated on the same complete set of scenes.
Table 1. Full-distribution evaluation on Monodepth2. Lower is better.
| 3D2Fool | 3D-APA | 3D-SSAP | RobCamou | |||||||||
|---|---|---|---|---|---|---|---|---|---|---|---|---|
| Method | ||||||||||||
| No Defense | 1.300 | 2.865 | 1.043 | 2.057 | 13.886 | 1.184 | 2.192 | 13.684 | 1.174 | 2.724 | 19.828 | 1.393 |
| Jedi | 1.269 | 2.025 | 1.117 | 1.220 | 2.564 | 1.143 | 1.304 | 2.910 | 1.145 | 1.402 | 4.041 | 1.194 |
| PAD | 1.582 | 1.389 | 1.581 | 1.623 | 2.845 | 1.560 | 1.689 | 3.939 | 1.571 | 1.981 | 6.909 | 1.674 |
| APEX | 1.488 | 2.552 | 1.329 | 1.940 | 8.124 | 1.387 | 2.178 | 9.393 | 1.407 | 2.241 | 10.231 | 1.521 |
| CamouPurifier$^\dagger$ | 1.260 | 1.471 | 1.269 | 1.366 | 1.526 | 1.379 | 1.345 | 1.512 | 1.358 | 1.399 | 1.639 | 1.399 |
| CamouPurifier$^\ddagger$ | 0.950 | 0.816 | 0.957 | 0.992 | 1.136 | 1.001 | 0.986 | 1.180 | 0.988 | 1.148 | 2.393 | 1.048 |
Key findings.
- Consistent effectiveness on the full distribution. CamouPurifier achieves lower vehicle-region discrepancies than all evaluated baseline defenses across all four attacks.
- Substantial attack suppression. Averaged across the four attacks, the vehicle-region discrepancy decreases from 12.566 without defense to 1.537 for masking-only purification and 1.381 for CamouPurifier$^\ddagger$, corresponding to reductions of 87.8% and 89.0%, respectively.
- Improved CamouPurifier$^\ddagger$ performance on the full set. The CamouPurifier$^\ddagger$ achieves the lowest average global and background discrepancies of 1.019 and 0.998, respectively. Its average vehicle-region discrepancy is also lower than that of masking-only purification on the complete evaluation set.
Author Response to Reviewer iCcG (Part 3/3)
3. Response to W3
Several claims are not fully supported by the current evaluation. First, although the paper reports clean-image depth discrepancy, it does not report the localization false-positive rate or predicted mask area on benign vehicles. Because both localization branches use percentile-based thresholds, it would be useful to evaluate whether naturally complex vehicle appearances, such as decals, reflections, or patterned paint, trigger unnecessary purification. Second, the adaptive attack is only briefly described, and the DINO percentile threshold is changed specifically for this experiment rather than fixed beforehand. More details about the adaptive objective and optimization procedure are needed. Finally, while the aggregate results are strong, the defense can degrade performance for some individual attack–model pairs. For example, under MIMDepth and 3D²Fool, the vehicle-region discrepancy increases from 2.860 without defense to 3.771 and 3.209 for the two variants. A clearer discussion of such failure cases would help qualify the overall conclusions.
We sincerely thank the reviewer for these constructive suggestions. We agree that a comprehensive evaluation should address false positives on benign vehicles, adaptive robustness, and individual cases where purification may degrade performance.
-
(1) False-positive mask rate (FPMR) We agree that clean-image depth discrepancy alone does not directly quantify unnecessary purification. Although the intersection of SAL and MTAL is intended to suppress spurious localization, percentile-based thresholding may still select regions on benign vehicles.
Table 2: FPMR on 1,043 clean vehicles. Precision, Recall, F1, and IoU are localization scores on the attacked-image.
FPMR Precision Recall F1 IoU Jedi 0.028 0.598 0.441 0.501 0.322 PAD 0.128 0.346 0.692 0.447 0.316 APEX 0.060 0.411 0.220 0.207 0.119 ours 0.148 0.630 0.708 0.652 0.495 Within the vehicle texture region, Jedi, PAD, APEX, and CamouPurifier mark 2.8%, 12.8%, 6.0%, and 14.8% of the pixels on average. On attacked images, the same CamouPurifier masks still give the highest Recall (0.708), F1 (0.652), and IoU (0.495).
In addition, we evaluated three types of normal graffiti: Graffiti-1 consists of localized colored graffiti; Graffiti-2 consists of global colored graffiti covering the entire vehicle body; and Graffiti-3 consists of global black-and-white graffiti covering the entire vehicle body. For each category, 1,024 images with random distances and angles were collected for further evaluation. is compared with the ground truth depth. CamouPurifier achieves strong adversarial camouflage localization, but its texture-sensitive design can also respond to visually complex benign vehicle appearances. PAD achieves the lowest error rate, but it completely blackens the vehicle region, thereby destroying the scene's semantic information.
Table 3: FPMR and on graffiti vehicles.
| Graffiti-1 | Graffiti-2 | Graffiti-3 | ||||
|---|---|---|---|---|---|---|
| FPMR | FPMR | FPMR | ||||
| no defense | - | 1.346 | - | 1.087 | - | 1.185 |
| Jedi | 0.082 | 1.92 | 0.374 | 2.045 | 0.227 | 1.613 |
| PAD | 0.353 | 1.52 | 0.952 | 1.144 | 0.673 | 1.26 |
| APEX | 0.077 | 2.082 | 0.108 | 2.681 | 0.149 | 2.694 |
| ours | 0.484 | 1.775 | 0.805 | 1.834 | 0.429 | 1.493 |
- (2) Failure cases and limitations. We thank the reviewer for highlighting the importance of analyzing individual attack–model combinations. However, we would like to clarify a minor misinterpretation of the results in Table 1. The reported values of 3.771 and 3.209 correspond to the full-image discrepancy , rather than the vehicle-region discrepancy . Specifically, under MIMDepth and 3D²Fool, the undefended vehicle-region discrepancy is 2.860. CamouPurifier† increases this value to 3.164, whereas CamouPurifier‡ reduces it to 2.573, achieving a 10.0% improvement over the undefended setting. Therefore, the complete reconstruction-based defense does improve vehicle-region depth consistency in this case. Nevertheless, we acknowledge that the full-image discrepancy increases from 3.134 without defense to 3.771 and 3.209 for the two variants, respectively. Our observations suggest that MIMDepth is particularly sensitive to localized appearance modifications, which can induce depth-prediction changes beyond the modified vehicle region.
Author Response to Reviewer iCcG (Part 1/3)
1. Response to W1
We thank the reviewer for highlighting the computational overhead of CamouPurifier. Our main contribution lies in proposing train-free, adversarial camouflage-agnostic and effective localization and purification of full-surface adversarial camouflage without retraining or modifying downstream MDE models. We acknowledge that the current implementation does not yet meet the stringent real-time requirements of autonomous driving, particularly when diffusion-based reconstruction is enabled.
CamouPurifier provides two configurations with different computational costs. The localization-only variant achieves 2.705 FPS, bringing a 27.6× speedup compared with the CamouPurifier$^\ddagger$ pipeline (0.098 FPS). The CamouPurifier$^\ddagger$ additionally aims to preserve visual and structural consistency through generative restoration. Notably, even the CamouPurifier$^\ddagger$ variant runs faster than the evaluated PAD (0.020 FPS) and APEX (0.026 FPS) implementations, showing its favorable computational efficiency among reconstruction-based defenses. All reported FPS measurements include only the defense purification stage and exclude downstream MDE inference.
We agree that further acceleration is necessary for real-time deployment. Potential directions include diffusion model distillation to reduce iterative sampling overhead and keyframe-based selective purification to avoid redundant processing of every video frame.