mlsanityData to 7 Oct 2026

IMPRESS: Evaluating the Resilience of Imperceptible Perturbations Against Unauthorized Data Usage in Diffusion-Based Generative AI

arXiv 2310.19248 ↗First posted 30 Oct 2023cs.CV cs.AI cs.CRcited by 32

On the map: Diffusion models → Robustness, privacy and adversarial aspects

Abstract (arXiv)

Diffusion-based image generation models, such as Stable Diffusion or DALL-E 2, are able to learn from given images and generate high-quality samples following the guidance from prompts. For instance, they can be used to create artistic images that mimic the style of an artist based on his/her original artworks or to maliciously edit the original images for fake content. However, such ability also brings serious ethical issues without proper authorization from the owner of the original images. In response, several attempts have been made to protect the original images from such unauthorized data usage by adding imperceptible perturbations, which are designed to mislead the diffusion model and make it unable to properly generate new samples. In this work, we introduce a perturbation purification platform, named IMPRESS, to evaluate the effectiveness of imperceptible perturbations as a protective measure. IMPRESS is based on the key observation that imperceptible perturbations could lead to a perceptible inconsistency between the original image and the diffusion-reconstructed image, which can be used to devise a new optimization strategy for purifying the image, which may weaken the protection of the original image from unauthorized data usage (e.g., style mimicking, malicious editing). The proposed IMPRESS platform offers a comprehensive evaluation of several contemporary protection methods, and can be used as an evaluation platform for future protection methods.

Key facts. Cited by 32 papers in the map (data through 7 Oct 2026). Placed in Robustness, privacy and adversarial aspects, one of 14 directions in Diffusion models. Builds on 39 of its 53 references that are in the map. First posted on arXiv 30 Oct 2023.

Builds on

39 of 53 references.

Closest ancestors · 2

Same direction, newest first

  1. 2023JPEG Compressed Images Can Bypass Protections Against AI Editing5 Apr
  2. 2023Raising the Cost of Malicious AI-Powered Image Editing13 Feb

Other references · 37

  1. 2021Learning Transferable Visual Models From Natural Language Supervisioncited by 33,456
  2. 2020Denoising Diffusion Probabilistic Modelscited by 20,851
  3. 2015U-Net: Convolutional Networks for Biomedical Image Segmentationcited by 19,604
  4. 2021High-Resolution Image Synthesis with Latent Diffusion Modelscited by 18,694
  5. 2018The Unreasonable Effectiveness of Deep Features as a Perceptual Metriccited by 11,029
  6. 2020Score-Based Generative Modeling through Stochastic Differential Equationscited by 8,964
  7. 2021Diffusion Models Beat GANs on Image Synthesiscited by 8,185
  8. 2017Towards Deep Learning Models Resistant to Adversarial Attackscited by 6,326
  9. 2018A Style-Based Generator Architecture for Generative Adversarial Networkscited by 5,986
  10. 2016Image-to-Image Translation with Conditional Adversarial Networkscited by 4,510
  11. 2021Improved Denoising Diffusion Probabilistic Modelscited by 3,736
  12. 2017Progressive Growing of GANs for Improved Quality, Stability, and Variationcited by 2,967
25 more
  1. 2016Improved Techniques for Training GANscited by 2,938
  2. 2018Large Scale GAN Training for High Fidelity Natural Image Synthesiscited by 2,609
  3. 2014Conditional Generative Adversarial Netscited by 2,152
  4. 2021SDEdit: Guided Image Synthesis and Editing with Stochastic Differential Equationscited by 1,903
  5. 2022RePaint: Inpainting using Denoising Diffusion Probabilistic Modelscited by 1,419
  6. 2021Palette: Image-to-Image Diffusion Modelscited by 1,296
  7. 2021Cascaded Diffusion Models for High Fidelity Image Generationcited by 1,155
  8. 2022Diffusion Models in Vision: A Surveycited by 1,008
  9. 2021Score-based Generative Modeling in Latent Spacecited by 634
  10. 2016StackGAN: Text to Photo-realistic Image Synthesis with Stacked Generative Adversarial Networkscited by 575
  11. 2023Your Diffusion Model is Secretly a Zero-Shot Classifiercited by 301
  12. 2020Uncovering the Limits of Adversarial Training against Norm-Bounded Adversarial Examplescited by 275
  13. 2023Glaze: Protecting Artists from Style Mimicry by Text-to-Image Modelscited by 200
  14. 2021Unlearnable Examples: Making Personal Data Unexploitablecited by 188
  15. 2015Large-scale Classification of Fine-Art Paintings: Learning The Right Metric on The Right Featurecited by 181
  16. 2021On Fast Sampling of Diffusion Probabilistic Modelscited by 170
  17. 2020Perceptual Adversarial Robustness: Defense Against Unseen Threat Modelscited by 137
  18. 2021Noise Estimation for Generative Diffusion Modelscited by 94
  19. 2021LowKey: Leveraging Adversarial Attacks to Protect Social Media Users from Facial Recognitioncited by 92
  20. 2022Text-Guided Synthesis of Artistic Images with Retrieval-Augmented Diffusion Modelscited by 63
  21. 2021Data Poisoning Won't Save You From Facial Recognitioncited by 45
  22. 2020Augmented Lagrangian Adversarial Attackscited by 28
  23. 2022DiVAE: Photorealistic Images Synthesis with Denoising Diffusion Decodercited by 27
  24. 2022Seeing is Living? Rethinking the Security of Facial Liveness Verification in the Deepfake Eracited by 26
  25. 2022The Swiss Army Knife for Image-to-Image Translation: Multi-Task Diffusion Modelscited by 24

Built on by

Cited by year (2026 to 7 Oct)

Citing papers per year: 2023: 2; 2024: 8; 2025: 13; 2026: 9.

32 papers.

Sort:

By area:

  1. 2023MetaCloak: Preventing Unauthorized Subject-driven Text-to-image Diffusion-based Synthesis via Meta-learningcited by 42
  2. 2024CDI: Copyrighted Data Identification in Diffusion Modelscited by 17
  3. 2025VideoEraser: Concept Erasure in Text-to-Video Diffusion Modelscited by 12
  4. 2024Replication in Visual Diffusion Models: A Survey and Outlookcited by 11
  5. 2025AdvPaint: Protecting Images from Inpainting Manipulation via Adversarial Attention Disruptioncited by 11
  6. 2024DiffVax: Optimization-Free Image Immunization Against Diffusion-Based Editingcited by 10
  7. 2024Dormant: Defending against Pose-driven Human Image Animationcited by 9
  8. 2025DCT-Shield: A Robust Frequency Domain Defense against Malicious Image Editingcited by 7
  9. 2023Large Scale Foundation Models for Intelligent Manufacturing Applications: A Surveycited by 6
  10. 2025Enhancing Variational Autoencoders with Smooth Robust Latent Encodingcited by 6
  11. 2025ArtistAuditor: Auditing Artist Style Pirate in Text-to-Image Generation Modelscited by 5
  12. 2025BlurGuard: A Simple Approach for Robustifying Image Protection Against AI-Powered Editingcited by 5
20 more
  1. 2026Anti-I2V: Safeguarding your photos from malicious image-to-video generationcited by 5
  2. 2024Rethinking and Red-Teaming Protective Perturbation in Personalized Diffusion Modelscited by 4
  3. 2025GuardDoor: Safeguarding Against Malicious Diffusion Editing via Protective Backdoorscited by 4
  4. 2025Beyond Vulnerabilities: A Survey of Adversarial Attacks as Both Threats and Defenses in Computer Vision Systemscited by 4
  5. 2024BridgePure: Limited Protection Leakage Can Break Black-Box Data Protectioncited by 3
  6. 2025CAT: Contrastive Adversarial Training for Evaluating the Robustness of Protective Perturbations in Latent Diffusion Modelscited by 3
  7. 2024Protective Perturbations against Unauthorized Data Usage in Diffusion-based Image Generationcited by 2
  8. 2025An h-space Based Adversarial Attack for Protection Against Few-shot Personalizationcited by 2
  9. 2025DIA: The Adversarial Exposure of Deterministic Inversion in Diffusion Modelscited by 2
  10. 2026Intellectual Property Protection for 3D Gaussian Splatting Assets: A Surveycited by 2
  11. 2026Bypassing Copyright Protection in Diffusion-based Customization via Two-Stage Latent Feature Optimizationcited by 2
  12. 2025StyleSentinel: Reliable Artistic Copyright Verification via Stylistic Fingerprintscited by 1
  13. 2026Purify Once, Edit Freely: Breaking Image Protections under Model Mismatchcited by 1
  14. 2024GLEAN: Generative Learning for Eliminating Adversarial Noisecited by 0
  15. 2025Beyond Known Fakes: Generalized Detection of AI-Generated Images via Post-hoc Distribution Alignmentcited by 0
  16. 2026SoK: Unlearnability and Unlearning for Model Dememorizationcited by 0
  17. 2026VOID: Defeating Unauthorized Mimicry in Latent Diffusion Modelscited by 0
  18. 2026Adversarial Attacks for Good: A Survey of Proactive Protection across the Visual Content Lifecyclecited by 0
  19. 2026IP Protection in the Era of Visual Generative AI: A Surveycited by 0
  20. 2026FeatMark: Feature-level Watermark Protection against Mimicry Attacks with Diffusion Modelscited by 0

Similar, not cited

Close in wording, no citation either way

  1. 2023Can Protective Perturbation Safeguard Personal Data from Being Exploited by Stable Diffusion?
  2. 2023Unlearnable Examples for Diffusion Models: Protect Data from Unauthorized Exploitation
  3. 2025Is Perturbation-Based Image Protection Disruptive to Image Editing?
  4. 2023EditShield: Protecting Unauthorized Image Editing by Instruction-guided Diffusion Models
  5. 2025Towards Robust Defense against Customization via Protective Perturbation Resistant to Diffusion-based Purification
  6. 2023Toward effective protection against diffusion based mimicry through score distillation
  7. 2024Nearly Zero-Cost Protection Against Mimicry by Personalized Diffusion Models
  8. 2024Imperceptible Protection against Style Imitation from Diffusion Models