Backdoor Attacks under Lossy Compression: From Failure to Reactivation and Adaptation
Listen
Radio episode about this paper
Transcript
Introduction to the show: ident: Security Radio. Generated commentary on the latest security and cryptography papers.
Nadia: I'm Nadia.
Elias: And I'm Elias, here with Priya.
Nadia: We're looking at the paper: "Backdoor Attacks under Lossy Compression".
Elias: Here's the gist of the paper. Real-world backdoor attacks often require poisoned datasets to be stored and transmitted before they are used to compromise deep learning systems.
Nadia: The paper highlights the necessity of explicitly accounting for lossy compression in backdoor attacks.
Elias: The core problem addressed is the failure of invisible triggers under lossy compression.
Nadia: Invisible triggers embedded in RGB images may not survive the conversion from RGB data to compressed binary bitstreams and back to decompressed RGB images.
Paper summary: Nadia: So wrapping up this discussion on "Backdoor Attacks under Lossy Compression: From Failure to Reactivation and Adaptation," the paper really hammers home that we need to explicitly account for lossy compression when designing data poisoning attacks in deep learning systems.
Elias: It moves the focus from just creating a trigger in an image to ensuring that trigger survives the whole storage and transmission pipeline, which is a crucial practical consideration >
Priya: The authors propose Universal Attack Reactivation and Compression-Adapted Attack as specific ways to deal with that, showing how we can reactivate degraded triggers or construct new ones adapted to the compression process >
Nadia: CAA is presented as a method that performs very well, maintaining high attack success rates and stealthiness even under JPEG re-compression, which suggests it's quite generalizable across different codecs >
Elias: It’s important to remember the limitation they acknowledge: they focus on revealing and constructing these compression aware backdoor attacks but don't propose a dedicated defense against these compression adapted threats >
Priya: So, while this work is great for understanding how these attacks work, the next step in research should be developing detection and mitigation methods specifically for backdoor attacks under lossy compression >
Conclusion: Nadia: So we’re wrapping up on "Backdoor Attacks under Lossy Compression," which is essentially about how invisible triggers can get ruined when you compress them for storage and transmission, and this paper shows two ways to fix that failure.
Elias: Yeah, they're proposing Universal Attack Reactivation and Compression-Adapted Attack as solutions to keep those triggers alive after the lossy compression process.
Priya: From what I see, it’s about making sure the trigger survives the journey from being created in a clean image to being used by a deep learning model. It gets into how those subtle high-frequency details that make an attack stealthy are exactly what get distorted during JPEG or similar compressions.
Nadia: Exactly, and what’s interesting is that they show these methods work across different types of image compression, not just one specific codec.
Elias: They demonstrate effectiveness with all-to-one attack success rates over four different backbone models, which suggests these strategies aren't just theoretical ideas; they actually hold up in practice.
Priya: But we have to watch the caveats here. They acknowledge that this work focuses on building and revealing these attacks, not necessarily creating a perfect defense against them yet.
Nadia: That’s fair, and the paper points out that while CAA is pretty robust against things like Fine-Pruning defenses, they haven't solved every possible way to stop these compression adapted threats.
Elias: The implication for us is that if you’re building systems where data gets compressed before training, you can’t just assume a trigger will survive; you have to design the attack or the defense around that lossy step.
Priya: And what this means for people on the ground is that if we use these methods to understand vulnerabilities, we need more research into how to actually detect these compromised models after they've been stored and recompressed.
Nadia: Right, so while the attacks are getting better at surviving compression, the next big challenge is building a system that can reliably flag a model poisoned this way.
Elias: Exactly. So now we’ve seen how to build the problem under these real-world conditions; next time, we need to figure out how to stop it in the wild.
Qian Li, Yunuo Chen, ⨂Yuntian Chen
Shanghai Jiao Tong University · Eastern Institute of Technology
cs.CR, cs.CV
Submitted: 2026-03-14
Updated: 2026-11-04
Comments: Accepted to NeurIPS 2026
License: http://creativecommons.org/licenses/by/4.0/
Importance score: 99/100
The gist: Real-world backdoor attacks often require poisoned datasets to be stored and transmitted before they are used to compromise deep learning systems.
Key concepts
- Lossy Compression Failure
- When an invisible backdoor is embedded in an image and then compressed (like JPEG), the resulting distorted data often loses the trigger's malicious information. This happens because lossy compression inevitably introduces errors, which can attenuate or completely destroy the subtle signals that activate the backdoor during inference.
- Universal Attack Reactivation (UAR)
- This strategy targets models that support pixel-level Region of Interest (ROI) coding. It works by measuring the difference between a clean and poisoned image to create ROI masks. These masks are used to selectively preserve or reactivate the trigger information that might have been degraded by compression.
- Compression-Adapted Attack (CAA)
- This is a new attack method that builds triggers intrinsically linked to the compression process. It uses customized ROI masks during encoding to strategically allocate bitrate, ensuring the malicious information is encoded in a way that survives both traditional codecs and learned image compression (LIC).
Terminology
Summary
Real-world backdoor attacks often require poisoned datasets to be stored and transmitted before they are used to compromise deep learning systems. p. 0 Backdoor Attacks under Lossy Compression: From
Failure to Reactivation and Adaptation
Qian Li1,2 Yunuo Chen1 Yuntian Chen2∗
1Shanghai Jiao Tong University 2Eastern Institute of Technology
Abstract
Real-world backdoor attacks often require poisoned datasets to be stored and transmitted before they are used to compromise deep learning systems. In the era of
big data, however, the inevitable use of lossy compression poses a fundamental
challenge to invisible backdoor attacks. Our contributions are summarized as follows:
• We reveal that previously effective invisible triggers can become ineffective after lossy
compression. This finding highlights the necessity of explicitly accounting for inevitable
storage-and-transmission compression in data poisoning-based backdoor attacks.Where the paper says it
Where the paper says it
The core problem addressed is the failure of invisible triggers under lossy compression. p. 0 Backdoor Attacks under Lossy Compression: From
Failure to Reactivation and Adaptation
Qian Li1,2 Yunuo Chen1 Yuntian Chen2∗
1Shanghai Jiao Tong University 2Eastern Institute of Technology
Abstract
Real-world backdoor attacks often require poisoned datasets to be stored and transmitted before they are used to compromise deep learning systems. In the era of
big data, however, the inevitable use of lossy compression poses a fundamental
challenge to invisible backdoor attacks.Where the paper says it
Invisible triggers embedded in RGB images may not survive the conversion from RGB data to compressed binary bitstreams and back to decompressed RGB images. p. 1 embedded in RGB images may not survive the conversion from RGB data to compressed binary
bitstreams and back to decompressed RGB images. Formally, many invisible attacks construct poisoned samples by directly modifying RGB images, e. p. 1 embedded in RGB images may not survive the conversion from RGB data to compressed binary
bitstreams and back to decompressed RGB images. Formally, many invisible attacks construct poisoned samples by directly modifying RGB images, e. ,
xp = x + δ ∈ R
3×H×W . Yet, once xp is lossily compressed into a binary bitstream yˆ ∈ {0, 1}
n,
such as a JPEG file, there is no guarantee that the malicious trigger information is preserved in
yˆ. Yet, once xp is lossily compressed into a binary bitstream yˆ ∈ {0, 1}
n,
such as a JPEG file, there is no guarantee that the malicious trigger information is preserved in
yˆ. After decompression, the recovered poisoned image xˆp ∈ R
3×H×W may no longer contain an
effective trigger. After decompression, the recovered poisoned image xˆp ∈ R
3×H×W may no longer contain an
effective trigger. Since lossy compression inevitably introduces distortion, i. Since lossy compression inevitably introduces distortion, i. , xˆ ̸= x, it can attenuate
or destroy trigger signals. , xˆ ̸= x, it can attenuate
or destroy trigger signals. As a result, poisoned data that are effective in the raw RGB domain may
lose their malicious functionality after storage and transmission, causing backdoor injection to fail. As a result, poisoned data that are effective in the raw RGB domain may
lose their malicious functionality after storage and transmission, causing backdoor injection to fail. This phenomenon affects a broad range of existing invisible attacks, including both invisible-trigger
methods and clean-label attacks [22, 23, 24, 25, 26, 27].Where the paper says it
Where the paper says it
Where the paper says it
Where the paper says it
Where the paper says it
Where the paper says it
Where the paper says it
Where the paper says it
The authors propose two ROI-based strategies to counteract this failure. p. 1 This provides a
natural interface for manipulating how trigger-related information is preserved or encoded during
compression, without modifying the standard decompression process used by victims. Following this principle, we propose two ROI-based poisoning strategies tailored to inevitable lossy
compression from failure to reactivation and adaptation.Where the paper says it
Building on the region-of-interest (ROI) coding mechanism in image compression, they propose two poisoning strategies tailored to inevitable lossy compression. p. 0 This requires attackers to ensure that transmitted binary bitstreams preserve malicious trigger information, such that effective triggers can be induced
after decompression. Building on the region-of-interest (ROI) coding mechanism
in image compression, we propose two poisoning strategies tailored to inevitable
lossy compression. Building on the region-of-interest (ROI) coding mechanism
in image compression, we propose two poisoning strategies tailored to inevitable
lossy compression. First, we introduce Universal Attack Reactivation, a general
method that uses sample-specific ROI masks to reactivate trigger information in
bitstreams for learned image compression (LIC). First, we introduce Universal Attack Reactivation, a general
method that uses sample-specific ROI masks to reactivate trigger information in
bitstreams for learned image compression (LIC). Second, we present CompressionAdapted Attack, a new attack strategy that employs customized ROI masks to
encode trigger information into bitstreams and applies to both traditional codecs
and LIC.Where the paper says it
Where the paper says it
Where the paper says it
Universal Attack Reactivation (UAR) focuses on preserving existing triggers. p. 1 The first, Universal Attack Reactivation,
aims to reactivate existing invisible attacks that become ineffective after compression by preserving
trigger-relevant information during learned image compression (LIC) [36, 37, 38, 39]. The second,
Compression-Adapted Attack (CAA), goes one step further by designing triggers that are inherently
adapted to the compression process, using customized ROI masks to encode malicious information
into compressed bitstreams.Where the paper says it
The first route, introduced in Sec. p. 4 We use this mechanism to design two complementary attack routes under lossy compression. The first route, introduced in Sec. The first route, introduced in Sec. 3, targets LIC models with pixel-level ROI support and reactivates
existing invisible triggers by preserving trigger-relevant information. This strategy is particularly suitable for LICs with pixel-level ROI
support, where fine-grained bitrate allocation can be controlled through sample-specific ROI masks. Given a clean image x and its poisoned counterpart xp generated by an existing attack, a direct way
to identify trigger-relevant regions is to measure their residuals as ROI masks:
Mres = NormMean|x − xp|
, (3)
where Mean(·) averages over RGB channels, and Norm(z) = z−min(z)
max(z)−min(z)
normalizes the mask
values.Where the paper says it
Where the paper says it
Where the paper says it
Compression-Adapted Attack (CAA) constructs triggers intrinsically coupled with the compression process. p. 5 Different from prior compression-aware attacks that treat compression artifacts
as triggers or exploit distributional gaps between compressed poisoned samples and uncompressed
benign samples, our goal is not to use compression merely as an additional trigger source. Instead,
we design a new trigger that is intrinsically coupled with the lossy storage-and-transmission
process.Where the paper says it
The second route, introduced in Sec. p. 4 3, targets LIC models with pixel-level ROI support and reactivates
existing invisible triggers by preserving trigger-relevant information. The second route, introduced
in Sec. The second route, introduced
in Sec. 4, targets both traditional codecs and LICs with region-level ROI support and constructs
compression-adapted triggers by encoding malicious information through customized ROI masks. Concretely, we use a customized ROI mask M∗
for poisoned samples and a uniform mask Muniform
for benign samples. The decompressed poisoned and benign samples are respectively given by
xˆp = DecoderDeQ(Q(Encoder(xp,M∗
)))
, (5)
xˆb = DecoderDeQ(Q(Encoder(xb,Muniform)))
. The decompressed poisoned and benign samples are respectively given by
xˆp = DecoderDeQ(Q(Encoder(xp,M∗
)))
, (5)
xˆb = DecoderDeQ(Q(Encoder(xb,Muniform)))
. (6)
Unlike conventional invisible attacks that add perturbations in the RGB domain, CAA encodes
malicious information through bitrate allocation during compression.Where the paper says it
Where the paper says it
Where the paper says it
Where the paper says it
Experimental results demonstrate the effectiveness of both strategies. p. 0 Second, we present CompressionAdapted Attack, a new attack strategy that employs customized ROI masks to
encode trigger information into bitstreams and applies to both traditional codecs
and LIC. Extensive experiments demonstrate the effectiveness of both strategies.Where the paper says it
Extensive experiments demonstrate the effectiveness of both strategies. p. 0 Second, we present CompressionAdapted Attack, a new attack strategy that employs customized ROI masks to
encode trigger information into bitstreams and applies to both traditional codecs
and LIC. Extensive experiments demonstrate the effectiveness of both strategies. We first evaluate whether UAR can restore existing invisible
backdoor attacks that fail after lossy compression. 1 ( ReA denotes the UAR-reactivated variant
of a baseline attack) reports all-to-one attack results on different datasets and backbones. Our method achieves ASR of over 99. 9% on all four backbone models across all
datasets, while the BA remains comparable to the clean model. In contrast, the performance of CAA depends on both the re-compression
quality factor and the selected ROI mask. When the JPEG quality factor is set to the default or
a higher value, and a relatively sparse ROI mask is adopted, CAA consistently achieves an ASR
above 95%.Where the paper says it
Where the paper says it
Where the paper says it
Where the paper says it
The paper concludes that the proposed methods are robust against various defenses. p. 7 Resistance to defense methods. We evaluate
CAA method against the common defense methods, including Fine-Pruning [82], STRIP [83],
and Gaussian noise and blur defenses.Where the paper says it
The results of resisting Fine-Pruning are shown in Fig. p. 7 The introduction of defense methods is detailed in the
supplementary material. • The results of resisting Fine-Pruning are
shown in Fig. • The results of resisting Fine-Pruning are
shown in Fig. 5, which indicate that FinePruning cannot completely eliminate the backdoor implanted by our method. We evaluate
CAA method against the common defense methods, including Fine-Pruning [82], STRIP [83],
and Gaussian noise and blur defenses. The introduction of defense methods is detailed in the
supplementary material. Resistance to defense methods. We evaluate
CAA method against the common defense methods, including Fine-Pruning [82], STRIP [83],
and Gaussian noise and blur defenses. Since prior invisible triggers fail under lossy compression, comparing stealth against those failed
methods is meaningless; therefore we compare CAA with the visible attack BadNet [8]. 10
demonstrates that CAA remains effective under lossy compression while offering superior stealth. When the JPEG quality factor is set to the default or
a higher value, and a relatively sparse ROI mask is adopted, CAA consistently achieves an ASR
above 95%. These results demonstrate that CAA offers substantially stronger robustness to practical
re-compression than non-adaptive invisible attacks. When the JPEG quality factor is set to the default or
a higher value, and a relatively sparse ROI mask is adopted, CAA consistently achieves an ASR
above 95%. These results demonstrate that CAA offers substantially stronger robustness to practical
re-compression than non-adaptive invisible attacks.Where the paper says it
Where the paper says it
Where the paper says it
Where the paper says it
Where the paper says it
Where the paper says it
Where the paper says it
Limitations and future directions are acknowledged. p. 8 Additional results on complexity, defenses, and robustness are provided in Appendices A–F. 5 Conclusion and Limitations
In this paper, we study backdoor attacks under inevitable lossy compression, a practical yet underexplored setting in real-world data storage and transmission pipelines.Where the paper says it
A limitation of this work is that we focus on revealing and constructing compression-aware backdoor attacks, but do not propose a dedicated defense against such compression-adapted threats. p. 8 Experiments across multiple datasets, backbones, and codecs demonstrate
the effectiveness of our methods. A limitation of this work is that we focus on revealing and constructing compression-aware backdoor attacks, but do not propose a dedicated defense against such
compression-adapted threats. Experiments across multiple datasets, backbones, and codecs demonstrate
the effectiveness of our methods. A limitation of this work is that we focus on revealing and constructing compression-aware backdoor attacks, but do not propose a dedicated defense against such
compression-adapted threats. A limitation of this work is that we focus on revealing and constructing compression-aware backdoor attacks, but do not propose a dedicated defense against such
compression-adapted threats. Developing effective detection and mitigation methods for backdoor
attacks under lossy compression remains an important direction for future research. A limitation of this work is that we focus on revealing and constructing compression-aware backdoor attacks, but do not propose a dedicated defense against such
compression-adapted threats. Developing effective detection and mitigation methods for backdoor
attacks under lossy compression remains an important direction for future research.Where the paper says it
Where the paper says it
Where the paper says it
Where the paper says it
The paper provides quantitative evidence of the proposed strategies. p. 1 This provides a
natural interface for manipulating how trigger-related information is preserved or encoded during
compression, without modifying the standard decompression process used by victims. Following this principle, we propose two ROI-based poisoning strategies tailored to inevitable lossy
compression from failure to reactivation and adaptation.Where the paper says it
Table 7 presents the performance on different codecs, showing that CAA outperforms the baseline methods in terms of both benign accuracy (BA) and attack success rate (ASR). p. 5 All images are resized to 64 × 64
and treated as raw images. We report Benign Accuracy (BA), Attack Success Rate (ASR) [75], and
Attack Stealthiness (AS) to demonstrate the effectiveness of each attack method. In contrast, the performance of CAA depends on both the re-compression
quality factor and the selected ROI mask. When the JPEG quality factor is set to the default or
a higher value, and a relatively sparse ROI mask is adopted, CAA consistently achieves an ASR
above 95%. Such frequency pattern serves
as a trigger correlated with the target class, enabling a successful backdoor. 10 shows that our method is imperceptible
by both objective metrics and human perception (human inspection test setting follows WaNet [24]). Our method redistributes bits according to the ROI mask, assigning more bits to light (high-priority) regions
while allocating fewer bits to dark regions. As
a result, the overall bit allocation is adjusted
spatially rather than uniformly increased.Where the paper says it
Where the paper says it
Where the paper says it
Where the paper says it
The paper provides detailed visualizations of the attack mechanism. p. 0 Building on the region-of-interest (ROI) coding mechanism
in image compression, we propose two poisoning strategies tailored to inevitable
lossy compression. First, we introduce Universal Attack Reactivation, a general
method that uses sample-specific ROI masks to reactivate trigger information in
bitstreams for learned image compression (LIC).Where the paper says it
Figure 7 is the visualization of CAA. p. 8 These results demonstrate that CAA offers substantially stronger robustness to practical
re-compression than non-adaptive invisible attacks. Origin Compression-Adapted Residual Allocation Map
(a) (b) (c) (d)
Origin Poisoned by CAA Residual ROI mask
Figure 7: Visualization of CAA. All reported experiments of CAA use an 8×8 checkerboard. 8 further analyzes our CAA method
driven special frequency pattern: we split each
image into four regions according to the ROI
mask and show the DCT spectrum for each region. 8 further analyzes our CAA method
driven special frequency pattern: we split each
image into four regions according to the ROI
mask and show the DCT spectrum for each region. The poisoned images exhibit noticeably
stronger high-frequency signals in the top-left
and bottom-right regions than in the top-right
and bottom-left.Where the paper says it
Where the paper says it
Where the paper says it
The paper compares CAA against existing compression-based methods. p. 2 While
such methods improve robustness against compression, they typically require the attacker to participate in or control the training process of a specific classification model, for example by modifying
the training objective. This differs from the dataset-only poisoning setting considered in this paper,
where the attacker only releases a poisoned dataset and the victim trains a model using an arbitrary
standard training pipeline.Where the paper says it
Table 15, Table 16, Table 17, Table 18 and Table 19 present the complete comparison results of our CAA and the baseline methods on JPEG2000 [49], BPG[50], Cheng-2020 [81], and ELIC [59]. p. 6 01↑) 100. 00
Table 1: Comparison of methods across multiple datasets; additional results are in the Appendix. 01↑) 100. 00
Table 1: Comparison of methods across multiple datasets; additional results are in the Appendix. All images are resized to 64 × 64
and treated as raw images. We report Benign Accuracy (BA), Attack Success Rate (ASR) [75], and
Attack Stealthiness (AS) to demonstrate the effectiveness of each attack method. All images are resized to 64 × 64
and treated as raw images. We report Benign Accuracy (BA), Attack Success Rate (ASR) [75], and
Attack Stealthiness (AS) to demonstrate the effectiveness of each attack method. Consequently, the results in Tab. 9 show that our
method does not noticeably increase file size,
or equivalently, the resulting bit budget.Where the paper says it
Where the paper says it
Where the paper says it
Where the paper says it
Where the paper says it
The paper summarizes the contribution of the two strategies. p. 8 5 Conclusion and Limitations
In this paper, we study backdoor attacks under inevitable lossy compression, a practical yet underexplored setting in real-world data storage and transmission pipelines. We show that existing invisible
backdoor attacks can become ineffective after compression, and propose two ROI-based strategies to
address this challenge.Where the paper says it
The contributions are summarized as follows:
• We reveal that previously effective invisible triggers can become ineffective after lossy compression.
• We propose Universal Attack Reactivation, a general ROI-based method for LICs that reactivates invisible triggers degraded by lossy compression. p. 2 This finding highlights the necessity of explicitly accounting for inevitable
storage-and-transmission compression in data poisoning-based backdoor attacks. • We propose Universal Attack Reactivation, a general ROI-based method for LICs that reactivates invisible triggers degraded by lossy compression.Where the paper says it
• We propose Compression-Adapted Attack (CAA), a new invisible backdoor attack for both LICs and traditional codecs. p. 2 Extensive experiments show that it
significantly improves the attack success rates of prior invisible attacks under compression. • We propose Compression-Adapted Attack (CAA), a new invisible backdoor attack for
both LICs and traditional codecs.Where the paper says it
The paper provides detailed experimental settings.
We evaluate backdoor attack methods on GTSRB [72], CIFAR-10 [73], and CelebA [74]. p. 5 1 Experimental Setting
Datasets and Metrics. We evaluate backdoor attack methods on GTSRB [72], CIFAR-10 [73], and
CelebA [74]. Classifiers. We use ResNet [76], Pre-activation ResNet [77], MobileNetV2 [78], and SENet [79]. 01, decayed by 10× every 100 epochs. We evaluate ROI-supported traditional codecs, including BPG [50] and JPEG 2000 [49],
and ROI-integrated LIC models, including Mbt-2018 [80], Cheng-2020 [81], and ELIC [59]. We evaluate ROI-supported traditional codecs, including BPG [50] and JPEG 2000 [49],
and ROI-integrated LIC models, including Mbt-2018 [80], Cheng-2020 [81], and ELIC [59]. We compare with three invisible attacks, WaNet [24], FTrojan [23], and BppAttack [22],
two clean-label attacks, SAA [26] and Narcissus [27], and two compression-based methods, ECA
[30] and CBA [2].Where the paper says it
Where the paper says it
Where the paper says it
Where the paper says it
The paper discusses the relationship between triggers and frequency characteristics. p. 1 Right: Removing high-frequency components from poisoned
test samples via Fourier Transform significantly reduces the ASR, suggesting that high-frequency information is critical for many invisible triggers. We observe that this failure is closely related to the frequency characteristics of invisible triggers.Where the paper says it
We observe that this failure is closely related to the frequency characteristics of invisible triggers. p. 1 Right: Removing high-frequency components from poisoned
test samples via Fourier Transform significantly reduces the ASR, suggesting that high-frequency information is critical for many invisible triggers. We observe that this failure is closely related to the frequency characteristics of invisible triggers. We observe that this failure is closely related to the frequency characteristics of invisible triggers. Inspired by prior studies on
frequency-domain backdoor behavior [28, 29],
we find that many invisible triggers rely on
subtle high-frequency components to achieve
stealthiness. Inspired by prior studies on
frequency-domain backdoor behavior [28, 29],
we find that many invisible triggers rely on
subtle high-frequency components to achieve
stealthiness. These components are visually
unobtrusive, but often play a crucial role in
triggering the backdoor.Where the paper says it
Where the paper says it
Where the paper says it
The paper details the robustness of CAA against specific adversarial defenses. p. 2 While
such methods improve robustness against compression, they typically require the attacker to participate in or control the training process of a specific classification model, for example by modifying
the training objective. This differs from the dataset-only poisoning setting considered in this paper,
where the attacker only releases a poisoned dataset and the victim trains a model using an arbitrary
standard training pipeline.Where the paper says it
In Tab. 10, we demonstrate that CAA remains effective under lossy compression while offering superior stealth compared to the visible attack BadNet [8]. p. 8 Since prior invisible triggers fail under lossy compression, comparing stealth against those failed
methods is meaningless; therefore we compare CAA with the visible attack BadNet [8]. 10
demonstrates that CAA remains effective under lossy compression while offering superior stealth. The introduction of defense methods is detailed in the
supplementary material. • The results of resisting Fine-Pruning are
shown in Fig. 6 indicate that the entropy
ranges of the clean model and the backdoored model trained on our poisoned samples, are very
similar. Our method is resistant to STRIP in compression scenarios. Our method is resistant to STRIP in compression scenarios. • After applying Gaussian blur (kernel:3) to decompressed test samples, we achieved a BA of 91. Left: Higher compression rates
introduce stronger distortion and substantially hinder
backdoor injection, indicating that lossy compression
can severely damage invisible triggers. Right: Removing high-frequency components from poisoned
test samples via Fourier Transform significantly reduces the ASR, suggesting that high-frequency information is critical for many invisible triggers.Where the paper says it
Where the paper says it
Where the paper says it
Where the paper says it
Where the paper says it
The paper concludes with a statement on the generalizability of CAA. p. 7 Our method consistently demonstrates minimal BA loss and high ASR close
to 100% across different codecs. This performance underscores the broad applicability and
generalizability of CAA.Where the paper says it
Our method consistently demonstrates minimal BA loss and high ASR close to 100% across different codecs. p. 7 ResNet18 [76] is employed as the classification model. Our method consistently demonstrates minimal BA loss and high ASR close
to 100% across different codecs. Our method consistently demonstrates minimal BA loss and high ASR close
to 100% across different codecs. This performance underscores the broad applicability and
generalizability of CAA. As
a result, the overall bit allocation is adjusted
spatially rather than uniformly increased. Consequently, the results in Tab. As
a result, the overall bit allocation is adjusted
spatially rather than uniformly increased. Consequently, the results in Tab. As
a result, the overall bit allocation is adjusted
spatially rather than uniformly increased. Consequently, the results in Tab. Consequently, the results in Tab. 9 show that our
method does not noticeably increase file size,
or equivalently, the resulting bit budget.Where the paper says it
Where the paper says it
Where the paper says it
Where the paper says it
Where the paper says it
Where the paper says it
The paper acknowledges its limitations. p. 8 Additional results on complexity, defenses, and robustness are provided in Appendices A–F. 5 Conclusion and Limitations
In this paper, we study backdoor attacks under inevitable lossy compression, a practical yet underexplored setting in real-world data storage and transmission pipelines.Where the paper says it
A limitation of this work is that we focus on revealing and constructing compression-aware backdoor attacks, but do not propose a dedicated defense against such compression-adapted threats. Developing effective detection and mitigation methods for backdoor attacks under lossy compression remains an important direction for future research.
The paper provides quantitative evaluation of the reactivation method. p. 8 000 -
BadNet 21. 4
Table 10: Quantitative evaluation of stealthiness.Where the paper says it
Table 4 shows that UAR mainly preserves and reactivates existing trigger information instead of creating a new mask-only trigger. p. 6 Removing both the trigger and Mfreq further reduces ASR to nearly zero. Thus, UAR mainly
preserves and reactivates existing trigger information instead of creating a new mask-only trigger. Ablation studies of the ROI mask. 4 verifies whether UAR reactivates original triggers rather
than introducing a new ROI-induced shortcut.Where the paper says it
Where the paper says it
The paper provides quantitative evaluation of the adaptation method. p. 8 000 -
BadNet 21. 4
Table 10: Quantitative evaluation of stealthiness.Where the paper says it
Table 11 shows that our CAA mislabels nearly 100% of poisoned images with only a 0. p. 7 Our
method outperforms alternatives across three
datasets. Our CAA mislabels nearly 100% of
poisoned images with only a 0. Our CAA mislabels nearly 100% of
poisoned images with only a 0. 06% loss in BA. Our method consistently demonstrates minimal BA loss and high ASR close
to 100% across different codecs. This performance underscores the broad applicability and
generalizability of CAA.Where the paper says it
Where the paper says it
Where the paper says it
The paper provides quantitative evaluation of the compression-adapted attack. p. 7 We visualize original images, compressed attack samples,
reactivated samples, and the corresponding ROI masks. 3 Evaluation of Compression-Adapted Attack
All-to-one attack.Where the paper says it
Table 15, Table 16, Table 17, Table 18 and Table 19 present the complete comparison results of our CAA and the baseline methods on JPEG2000 [49], BPG[50], Cheng-2020 [81], and ELIC [59]. CAA outperforms the baseline methods in terms of both benign accuracy (BA) and attack success rate (ASR) across these codecs. The performance on different compression models consistently demonstrates minimal BA loss and high ASR close to 100%. p. 7 Our method consistently demonstrates minimal BA loss and high ASR close
to 100% across different codecs. This performance underscores the broad applicability and
generalizability of CAA.Where the paper says it
The paper provides a comprehensive comparison across multiple backbones. p. 6 01↑) 100. 00
Table 1: Comparison of methods across multiple datasets; additional results are in the Appendix.Where the paper says it
Table 16 shows that CAA outperforms the baseline methods in terms of both benign accuracy (BA) and attack success rate (ASR) across different backbones. p. 5 All images are resized to 64 × 64
and treated as raw images. We report Benign Accuracy (BA), Attack Success Rate (ASR) [75], and
Attack Stealthiness (AS) to demonstrate the effectiveness of each attack method. 01↑) 100. 00
Table 1: Comparison of methods across multiple datasets; additional results are in the Appendix.Where the paper says it
Where the paper says it
The paper provides a summary of the findings. p. 1 ROI coding allows codecs to assign different bitrates to different
image regions, thereby controlling the spatial distribution of compression distortion. This provides a
natural interface for manipulating how trigger-related information is preserved or encoded during
compression, without modifying the standard decompression process used by victims.Where the paper says it
In this paper, we study backdoor attacks under inevitable lossy compression, a practical yet underexplored setting in real-world data storage and transmission pipelines. p. 8 Additional results on complexity, defenses, and robustness are provided in Appendices A–F. 5 Conclusion and Limitations
In this paper, we study backdoor attacks under inevitable lossy compression, a practical yet underexplored setting in real-world data storage and transmission pipelines. 5 Conclusion and Limitations
In this paper, we study backdoor attacks under inevitable lossy compression, a practical yet underexplored setting in real-world data storage and transmission pipelines. We show that existing invisible
backdoor attacks can become ineffective after compression, and propose two ROI-based strategies to
address this challenge. We show that existing invisible
backdoor attacks can become ineffective after compression, and propose two ROI-based strategies to
address this challenge. Experiments across multiple datasets, backbones, and codecs demonstrate
the effectiveness of our methods.Where the paper says it
Where the paper says it
Where the paper says it
The paper provides a detailed overview of the pipeline. p. 0 Prior compression-based attacks typically exploit compression artifacts as certain triggers to distinguish poisoned RGB samples from
uncompressed benign ones, rather than addressing whether malicious information
can survive a shared lossy storage-and-transmission pipeline. In this paper, we
highlight the necessity of explicitly accounting for lossy compression in backdoor
attacks.Where the paper says it
Figure 2 illustrates the dataset-only poisoning pipeline considered in this paper. p. 3 1 Threat Model and Problem Setup under Lossy Compression
Fig. 2 illustrates the dataset-only poisoning pipeline considered in this paper. Red borders denote poisoned
samples, and green borders denote benign samples. Column 1 shows the real-world dataset-only
poisoning pipeline, where data are stored and transmitted as compressed bitstreams before victimside training. Column 1 shows the real-world dataset-only
poisoning pipeline, where data are stored and transmitted as compressed bitstreams before victimside training. Column 2 illustrates that previous invisible attacks may fail after lossy compression
because trigger information is destroyed. Column 2 illustrates that previous invisible attacks may fail after lossy compression
because trigger information is destroyed. Columns 3-4 present our two ROI-based strategies:
Universal Attack Reactivation, which reactivates ineffective invisible attacks by preserving triggerrelevant information, and Compression-Adapted Attack, which encodes malicious information into
bitstreams to construct compression-adapted triggers for both traditional codecs and LICs.Where the paper says it
Where the paper says it
Where the paper says it
Where the paper says it
The paper provides a detailed overview of the ROI coding mechanism. p. 1 ROI coding allows codecs to assign different bitrates to different
image regions, thereby controlling the spatial distribution of compression distortion. This provides a
natural interface for manipulating how trigger-related information is preserved or encoded during
compression, without modifying the standard decompression process used by victims.Where the paper says it
ROI coding allows codecs to assign different bitrates to different image regions, thereby controlling the spatial distribution of compression distortion. p. 1 We achieve this by leveraging the Region of Interest (ROI) coding mechanism in image
compression [31, 32, 33, 34, 35]. ROI coding allows codecs to assign different bitrates to different
image regions, thereby controlling the spatial distribution of compression distortion. This interface is well suited to our goal. Instead of introducing additional backdoor modules or
requiring a modified decoder, ROI masks allow the attacker to guide the standard compression process
itself. 69, 32, 33]. An ROI-supported encoder takes an additional mask M = [mij ] ∈ R
H×W as input:
Compression : yˆ = Q(Encoder(x,M)).Where the paper says it
Where the paper says it
Where the paper says it
The paper discusses the relationship between compression and trigger information. p. 0 In this paper, we
highlight the necessity of explicitly accounting for lossy compression in backdoor
attacks. This requires attackers to ensure that transmitted binary bitstreams preserve malicious trigger information, such that effective triggers can be induced
after decompression.Where the paper says it
Stronger compression substantially reduces the attack success rate, and explicitly removing high-frequency components from poisoned test samples also leads to a significant ASR drop. p. 1 As shown in Fig. 1,
stronger compression substantially reduces the
attack success rate, and explicitly removing
high-frequency components from poisoned
test samples also leads to a significant ASR
drop. 1,
stronger compression substantially reduces the
attack success rate, and explicitly removing
high-frequency components from poisoned
test samples also leads to a significant ASR
drop. These observations suggest that invisible triggers achieve stealth by hiding in
compression-sensitive high-frequency information, while lossy compression tends to
suppress or distort exactly such information.Where the paper says it
Where the paper says it
The paper provides a final statement on the study's scope.
We consider storage and transmission with publicly available codec implementations or specifications, so the attack is evaluated under realistic compression pipelines. p. 3 The attacker aims to release a poisoned dataset whose decompressed samples can
successfully implant a backdoor into a victim-trained model, while maintaining normal accuracy on
decompressed benign samples. We consider storage and transmission with publicly available codec
implementations or specifications, so the attack is evaluated under realistic compression pipelines. In
contrast, they do not explicitly address whether malicious information can survive a shared lossy
storage-and-transmission pipeline, where both benign and poisoned data are compressed into binary
bitstreams. Our goal is therefore not merely to use compression artifacts as another form of trigger. Our goal is therefore not merely to use compression artifacts as another form of trigger. Instead, we study backdoor attacks under inevitable lossy compression, where the attacker must ensure
that the transmitted bitstream recovers or induces malicious trigger information after decompression. This introduces a fundamental challenge to invisible backdoor attacks: triggers
∗Corresponding Author
Preprint. arXiv:2603. arXiv:2603. 13864v2 [cs. 13864v2 [cs. CR] 15 May 2026Where the paper says it
Where the paper says it
Where the paper says it
Where the paper says it
Where the paper says it
Where the paper says it
The paper details the specific attack strategies. p. 0 Building on the region-of-interest (ROI) coding mechanism
in image compression, we propose two poisoning strategies tailored to inevitable
lossy compression. First, we introduce Universal Attack Reactivation, a general
method that uses sample-specific ROI masks to reactivate trigger information in
bitstreams for learned image compression (LIC).Where the paper says it
To address this challenge, we explicitly incorporate lossy compression into the design of backdoor attacks. p. 1 Instead, we study backdoor attacks under inevitable lossy compression, where the attacker must ensure
that the transmitted bitstream recovers or induces malicious trigger information after decompression. To address this challenge, we explicitly incorporate lossy compression into the design of backdoor
attacks. Instead of introducing additional backdoor modules or
requiring a modified decoder, ROI masks allow the attacker to guide the standard compression process
itself. We use this mechanism to design two complementary attack routes under lossy compression. The first route, introduced in Sec. 3, targets LIC models with pixel-level ROI support and reactivates
existing invisible triggers by preserving trigger-relevant information. The second route, introduced
in Sec. 4, targets both traditional codecs and LICs with region-level ROI support and constructs
compression-adapted triggers by encoding malicious information through customized ROI masks.Where the paper says it
Where the paper says it
Where the paper says it
Where the paper says it
The paper provides a final statement on resilience. p. 7 86% and ASR reached 100%. These results demonstrate our method’s resilience against such defenses.Where the paper says it
CAA offers substantially stronger robustness to practical re-compression than non-adaptive invisible attacks. The results in Tab. 8 report the ASR under JPEG re-compression, where CAA consistently achieves an ASR above 95% when the JPEG quality factor is set to the default or a higher value and a relatively sparse ROI mask is adopted. p. 8 In contrast, the performance of CAA depends on both the re-compression
quality factor and the selected ROI mask. When the JPEG quality factor is set to the default or
a higher value, and a relatively sparse ROI mask is adopted, CAA consistently achieves an ASR
above 95%. Since FTrojan is not designed to be compressionadaptive, its trigger remains highly vulnerable to re-compression, even when it is injected before the
initial compression stage. In contrast, the performance of CAA depends on both the re-compression
quality factor and the selected ROI mask. 8 reports the ASR under JPEG re-compression. As
expected, FTrojanReA almost completely fails.Where the paper says it
Where the paper says it
Where the paper says it
Green — the paper says it, and the page is printed. Red — no passage carries it, so treat that sentence as unverified. The same check runs on your own text: check a summary against its source, $5.
Improvements for AI systems
-
Universal Attack Reactivation (UAR) improves attack success rates of prior invisible attacks under compression by
preserving trigger-relevant information during learned image compression (LIC)
. This allows for the reactivation of triggers that otherwise become ineffective after lossy compression. -
Compression-Adapted Attack (CAA) constructs triggers that are
intrinsically coupled with the lossy storage-and-transmission process
by using ROI masks to encode malicious information into compressed bitstreams. This induces adistinctive yet visually imperceptible frequency-distribution pattern in decompressed poisoned images
which serves as a compression-adapted trigger.
Abstract
Real-world backdoor attacks often require poisoned datasets to be stored and transmitted before they are used to compromise deep learning systems. In the era of big data, however, the inevitable use of lossy compression poses a fundamental challenge to invisible backdoor attacks. We observe that triggers embedded in RGB images can become ineffective once the images are lossily compressed into binary bitstreams, such as JPEG files, for storage and transmission. Consequently, poisoned data may lose their malicious functionality after compression, causing backdoor injection to fail. Prior compression-based attacks typically exploit compression artifacts as certain triggers to distinguish poisoned RGB samples from uncompressed benign ones, rather than addressing whether malicious information can survive a shared lossy storage-and-transmission pipeline. In this paper, we highlight the necessity of explicitly accounting for lossy compression in backdoor attacks. This requires attackers to ensure that transmitted binary bitstreams preserve malicious trigger information, such that effective triggers can be induced after decompression. Building on the region-of-interest (ROI) coding mechanism in image compression, we propose two poisoning strategies tailored to inevitable lossy compression. First, we introduce Universal Attack Reactivation, a general method that uses sample-specific ROI masks to reactivate trigger information in bitstreams for learned image compression (LIC). Second, we present Compression-Adapted Attack, a new attack strategy that employs customized ROI masks to encode trigger information into bitstreams and applies to both traditional codecs and LIC. Extensive experiments demonstrate the effectiveness of both strategies.
Sources
- BadNets: Identifying Vulnerabilities in the Machine Learning Model Supply Chain
- Everyone Can Attack: Repurpose Lossy Compression as a Natural Backdoor Attack
- S2CFormer: Revisiting the RD-Latency Trade-off in Transformer-based Learned Image Compression
- Targeted Backdoor Attacks on Deep Learning Systems Using Data Poisoning
- Label-Consistent Backdoor Attacks
- Learned Image Compression with Hierarchical Progressive Context Modeling
- MambaIC: State Space Models for High-Performance Learned Image Compression
- NNoculation: Catching BadNets in the Wild
Related papers
- From Strategy Search to Policy Learning: MCTS-Guided Adversarial Co-Evolution for LLM Safety Alignment
- Privacy Leakage from Recurrent State in LLM Inference
- Incrementally Verifiable Oracle Aided Computation
- Privilege and confidentiality in generative AI workflows
- BadMoE: Backdooring Mixture-of-Experts LLMs via Routing-Aware Triggers and Task-Underutilized Experts
- CoVault: Secure, Scalable Analytics of Personal Data