mlsanityData to 7 Oct 2026

AEROBLADE: Training-Free Detection of Latent Diffusion Images Using Autoencoder Reconstruction Error

arXiv 2401.17879 ↗First posted 31 Jan 2024cs.CVcited by 123

On the map: Diffusion models → Image generation, text-to-image and editing

Abstract (arXiv)

With recent text-to-image models, anyone can generate deceptively realistic images with arbitrary contents, fueling the growing threat of visual disinformation. A key enabler for generating high-resolution images with low computational cost has been the development of latent diffusion models (LDMs). In contrast to conventional diffusion models, LDMs perform the denoising process in the low-dimensional latent space of a pre-trained autoencoder (AE) instead of the high-dimensional image space. Despite their relevance, the forensic analysis of LDMs is still in its infancy. In this work we propose AEROBLADE, a novel detection method which exploits an inherent component of LDMs: the AE used to transform images between image and latent space. We find that generated images can be more accurately reconstructed by the AE than real images, allowing for a simple detection approach based on the reconstruction error. Most importantly, our method is easy to implement and does not require any training, yet nearly matches the performance of detectors that rely on extensive training. We empirically demonstrate that AEROBLADE is effective against state-of-the-art LDMs, including Stable Diffusion and Midjourney. Beyond detection, our approach allows for the qualitative analysis of images, which can be leveraged for identifying inpainted regions. We release our code and data at https://github.com/jonasricker/aeroblade .

Key facts. Cited by 123 papers in the map (data through 7 Oct 2026). Placed in Image generation, text-to-image and editing, one of 14 directions in Diffusion models. Builds on 46 of its 62 references that are in the map. First posted on arXiv 31 Jan 2024.

Builds on

46 of 62 references.

Closest ancestors · 8

Same direction, newest first

  1. 2023Kandinsky: an Improved Text-to-Image Synthesis with Image Prior and Latent Diffusion5 Oct
  2. 2023Exposing the Fake: Effective Diffusion-Generated Images Detection12 Jul
  3. 2023Intriguing properties of synthetic images: from generative adversarial networks to diffusion models13 Apr
  4. 2023DIRE for Diffusion-Generated Image Detection16 Mar
  5. 2022On the detection of synthetic images generated by diffusion models1 Nov
  6. 2022Towards the Detection of Diffusion Model Deepfakes26 Oct
  7. 2022DE-FAKE: Detection and Attribution of Fake Images Generated by Text-to-Image Generation Models13 Oct
  8. 2022Hierarchical Text-Conditional Image Generation with CLIP Latents13 Apr

Other references · 38

  1. 2021Learning Transferable Visual Models From Natural Language Supervisioncited by 33,456
  2. 2020An Image is Worth 16x16 Words: Transformers for Image Recognition at Scalecited by 29,915
  3. 2020Denoising Diffusion Probabilistic Modelscited by 20,851
  4. 2015U-Net: Convolutional Networks for Biomedical Image Segmentationcited by 19,604
  5. 2021High-Resolution Image Synthesis with Latent Diffusion Modelscited by 18,694
  6. 2014Very Deep Convolutional Networks for Large-Scale Image Recognitioncited by 15,267
  7. 2018The Unreasonable Effectiveness of Deep Features as a Perceptual Metriccited by 11,029
  8. 2020Denoising Diffusion Implicit Modelscited by 9,655
  9. 2014ImageNet Large Scale Visual Recognition Challengecited by 8,441
  10. 2021Diffusion Models Beat GANs on Image Synthesiscited by 8,185
  11. 2015Deep Unsupervised Learning using Nonequilibrium Thermodynamicscited by 7,005
  12. 2017Neural Discrete Representation Learningcited by 4,650
26 more
  1. 2022LAION-5B: An open large-scale dataset for training next generation image-text modelscited by 4,131
  2. 2019Generative Modeling by Estimating Gradients of the Data Distributioncited by 4,067
  3. 2013Auto-Encoding Variational Bayescited by 3,980
  4. 2021Align before Fuse: Vision and Language Representation Learning with Momentum Distillationcited by 1,612
  5. 2022Diffusion Models: A Comprehensive Survey of Methods and Applicationscited by 1,107
  6. 2019FaceForensics++: Learning to Detect Manipulated Facial Imagescited by 1,095
  7. 2016SqueezeNet: AlexNet-level accuracy with 50x fewer parameters and <0.5MB model sizecited by 1,054
  8. 2019CNN-generated images are surprisingly easy to spot... for nowcited by 842
  9. 2020Image Quality Assessment: Unifying Structure and Texture Similaritycited by 792
  10. 2020Leveraging Frequency Analysis for Deep Fake Image Recognitioncited by 501
  11. 2023Towards Universal Fake Image Detectors that Generalize Across Generative Modelscited by 483
  12. 2021GAN Inversion: A Surveycited by 366
  13. 2023The Stable Signature: Rooting Watermarks in Latent Diffusion Modelscited by 289
  14. 2022Diffusion Models for Medical Anomaly Detectioncited by 231
  15. 2020Artificial Fingerprinting for Generative Models: Rooting Deepfake Attribution in Training Datacited by 200
  16. 2021Are GAN generated images easy to detect? A critical analysis of the state-of-the-artcited by 139
  17. 2020Responsible Disclosure of Generative Models Using Scalable Fingerprintingcited by 90
  18. 2021A Closer Look at Fourier Spectrum Discrepancies for CNN-generated Images Detectioncited by 68
  19. 2021On the Frequency Bias of Generative Modelscited by 60
  20. 2020Exposing GAN-generated Faces Using Inconsistent Corneal Specular Highlightscited by 59
  21. 2023Unsupervised Out-of-Distribution Detection with Diffusion Inpaintingcited by 57
  22. 2023Anomaly Detection with Conditioned Denoising Diffusion Modelscited by 56
  23. 2021Eyes Tell All: Irregular Pupil Shapes Reveal GAN-generated Facescited by 48
  24. 2022Lighting (In)consistency of Paint by Textcited by 35
  25. 2021Towards Universal GAN Image Detectioncited by 26
  26. 2023Exploring Diffusion Models for Unsupervised Video Anomaly Detectioncited by 23

Built on by

Cited by year (2026 to 7 Oct)

Citing papers per year: 2024: 24; 2025: 44; 2026: 55.

123 papers.

Sort:

By area:

  1. 2024A Sanity Check for AI-generated Image Detectioncited by 187
  2. 2024Any-Resolution AI-Generated Image Detection by Spectral Learningcited by 56
  3. 2024Zero-Shot Detection of AI-Generated Imagescited by 50
  4. 2024Community Forensics: Using Thousands of Generators to Train Fake Image Detectorscited by 50
  5. 2024FIRE: Robust Detection of Diffusion-Generated Images via Frequency-Guided Reconstruction Errorcited by 46
  6. 2024On Learning Multi-Modal Forgery Representation for Diffusion Generated Video Detectioncited by 40
  7. 2024FakeBench: Probing Explainable Fake Image Detection via Large Multimodal Modelscited by 36
  8. 2024RIGID: A Training-free and Model-Agnostic Framework for Robust AI-Generated Image Detectioncited by 36
  9. 2024Aligned Datasets Improve Detection of Latent Diffusion-Generated Imagescited by 36
  10. 2025Bridging the Gap Between Ideal and Real-world Evaluation: Benchmarking AI-Generated Image Detection in Challenging Scenarioscited by 24
  11. 2024Circumventing shortcuts in audio-visual deepfake detection datasets with unsupervised learningcited by 23
  12. 2024Are CLIP features all you need for Universal Synthetic Image Origin Attribution?cited by 14
38 more
  1. 2025Few-Shot Learner Generalizes Across AI-Generated Image Detectioncited by 13
  2. 2024A Survey of Defenses Against AI-Generated Visual Media: Detection,Disruption, and Authenticationcited by 12
  3. 2024Vulnerabilities in AI-generated Image Detection: The Challenge of Adversarial Attackscited by 12
  4. 2025Survey on AI-Generated Media Detection: From Non-MLLM to MLLMcited by 12
  5. 2025Methods and Trends in Detecting AI-Generated Images: A Comprehensive Reviewcited by 12
  6. 2025Stay-Positive: A Case for Ignoring Real Image Features in Fake Image Detectioncited by 11
  7. 2025Forensic Self-Descriptions Are All You Need for Zero-Shot Detection, Open-Set Source Attribution, and Clustering of AI-generated Imagescited by 11
  8. 2025FakeScope: Large Multimodal Expert Model for Transparent AI-Generated Image Forensicscited by 11
  9. 2024AI-Generated Faces in the Real World: A Large-Scale Case Study of Twitter Profile Imagescited by 10
  10. 2024Deepfake Media Generation and Detection in the Generative AI Era: A Survey and Outlookcited by 10
  11. 2025Manifold Induced Biases for Zero-shot and Few-shot Detection of Generated Imagescited by 10
  12. 2025AI-GenBench: A New Ongoing Benchmark for AI-Generated Image Detectioncited by 10
  13. 2026DNA: Uncovering Universal Latent Forgery Knowledgecited by 10
  14. 2024HFI: A unified framework for training-free detection and implicit watermarking of latent diffusion model generated imagescited by 9
  15. 2025Training-free Detection of AI-generated images via Cropping Robustnesscited by 9
  16. 2025Towards Explainable Fake Image Detection with Multi-Modal Large Language Modelscited by 8
  17. 2025Navigating the Challenges of AI-Generated Image Detection in the Wild: What Truly Matters?cited by 8
  18. 2026Training-free Detection of Generated Videos via Spatial-Temporal Likelihoodscited by 7
  19. 2024Adversarial Robustness of AI-Generated Image Detectors in the Real Worldcited by 6
  20. 2025Toward Medical Deepfake Detection: A Comprehensive Dataset and Novel Methodcited by 6
  21. 2025Self-Supervised AI-Generated Image Detection: A Camera Metadata Perspectivecited by 6
  22. 2025LATTE: Latent Trajectory Embedding for Diffusion-Generated Image Detectioncited by 5
  23. 2025AEDR: Training-Free AI-Generated Image Attribution via Autoencoder Double-Reconstructioncited by 5
  24. 2025Detecting Generated Images by Fitting Natural Image Distributionscited by 5
  25. 2026AIForge-Doc: A Benchmark for Detecting AI-Forged Tampering in Financial and Form Documentscited by 5
  26. 2024The Cat and Mouse Game: The Ongoing Arms Race Between Diffusion Models and Detection Methodscited by 4
  27. 2025Fully AI-Generated Image Detection: Definition, Recent Advances and Challengescited by 4
  28. 2025RAID: A Dataset for Testing the Adversarial Robustness of AI-Generated Image Detectorscited by 4
  29. 2025Towards Reliable Identification of Diffusion-based Image Manipulationscited by 4
  30. 2025Bi-Level Optimization for Self-Supervised AI-Generated Face Detectioncited by 4
  31. 2026Layer Consistency Matters: Elegant Latent Transition Discrepancy for Generalizable Synthetic Image Detectioncited by 4
  32. 2024A Large-scale Universal Evaluation Benchmark For Face Forgery Detectioncited by 3
  33. 2024Epistemic Uncertainty for Generated Image Detectioncited by 3
  34. 2025Model Discrepancy Learning: Synthetic Faces Detection Based on Multi-Reconstructioncited by 3
  35. 2025bfD3QE: Learning Discrete Distribution Discrepancy-aware Quantization Error for Autoregressive-Generated Image Detectioncited by 3
  36. 2025HandEval: Taking the First Step Towards Hand Quality Evaluation in Generated Imagescited by 3
  37. 2025RobustSora: De-Watermarked Benchmark for Robust AI-Generated Video Detectioncited by 3
  38. 2026Attribution as Retrieval: Model-Agnostic AI-Generated Image Attributioncited by 3

50 most cited of 123 All 123 (JSON)

Similar, not cited

Close in wording, no citation either way

  1. 2026LDM-is-AE: Latent Diffusion Model is an Auto-Encoder for End-to-End Image Generation
  2. 2026R2BD: A Reconstruction-Based Method for Generalizable and Efficient Detection of Fake Images
  3. 2026FIND: A Simple yet Effective Baseline for Diffusion-Generated Image Detection
  4. 2023Diffusion Noise Feature: Accurate and Fast Generated Image Detection
  5. 2025Detecting AI-Generated Images via Diffusion Snap-Back Reconstruction: A Forensic Approach
  6. 2024LaRE2: Latent Reconstruction Error Based Method for Diffusion-Generated Image Detection
  7. 2026FlowGuard: Towards Lightweight In-Generation Safety Detection for Diffusion Models via Linear Latent Decoding
  8. 2025DRAGON: A Large-Scale Dataset of Realistic Images Generated by Diffusion Models