mlsanityData to 14 Oct 2026

Towards Reliable Identification of Diffusion-based Image Manipulations

arXiv 2506.05466 ↗First posted 5 Jun 2025cs.CVcited by 4

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

Abstract (arXiv)

Changing facial expressions, gestures, or background details may dramatically alter the meaning conveyed by an image. Notably, recent advances in diffusion models greatly improve the quality of image manipulation while also opening the door to misuse. Identifying changes made to authentic images, thus, becomes an important task, constantly challenged by new diffusion-based editing tools. To this end, we propose a novel approach for ReliAble iDentification of inpainted AReas (RADAR). RADAR builds on existing foundation models and combines features from different image modalities. It also incorporates an auxiliary contrastive loss that helps to isolate manipulated image patches. We demonstrate these techniques to significantly improve both the accuracy of our method and its generalisation to a large number of diffusion models. To support realistic evaluation, we further introduce BBC-PAIR, a new comprehensive benchmark, with images tampered by 28 diffusion models. Our experiments show that RADAR achieves excellent results, outperforming the state-of-the-art in detecting and localising image edits made by both seen and unseen diffusion models. Our code, data and models will be publicly available at https://alex-costanzino.github.io/radar/.

Key facts. Cited by 4 papers in the map (data through 14 Oct 2026). Placed in Image generation, text-to-image and editing, one of 14 directions in Diffusion models. Builds on 61 of its 93 references that are in the map. First posted on arXiv 5 Jun 2025.

Builds on

61 of 93 references.

Closest ancestors · 13

Same direction, newest first

  1. 2024BrushEdit: All-In-One Image Inpainting and Editing13 Dec
  2. 2024Time Step Generating: A Universal Synthesized Deepfake Image Detector17 Nov
  3. 2024Kandinsky 3: Text-to-Image Synthesis for Multifunctional Generative Framework28 Oct
  4. 2024FakeInversion: Learning to Detect Images from Unseen Text-to-Image Models by Inverting Stable Diffusion12 Jun
  5. 2024Diffusion Model-Based Image Editing: A Survey27 Feb
  6. 2024AEROBLADE: Training-Free Detection of Latent Diffusion Images Using Autoencoder Reconstruction Error31 Jan
  7. 2023Kandinsky: an Improved Text-to-Image Synthesis with Image Prior and Latent Diffusion5 Oct
  8. 2023DIRE for Diffusion-Generated Image Detection16 Mar
  9. 2022Imagen Editor and EditBench: Advancing and Evaluating Text-Guided Image Inpainting13 Dec
  10. 2022Blended Latent Diffusion6 Jun
  11. 2022Photorealistic Text-to-Image Diffusion Models with Deep Language Understanding23 May
  12. 2021GLIDE: Towards Photorealistic Image Generation and Editing with Text-Guided Diffusion Models20 Dec
1 more
  1. 2021Blended Diffusion for Text-driven Editing of Natural Images29 Nov

Other references · 48

  1. 2014Adam: A Method for Stochastic Optimizationcited by 37,240
  2. 2021High-Resolution Image Synthesis with Latent Diffusion Modelscited by 18,863
  3. 2020A Simple Framework for Contrastive Learning of Visual Representationscited by 12,049
  4. 2023DINOv2: Learning Robust Visual Features without Supervisioncited by 7,974
  5. 2021Emerging Properties in Self-Supervised Vision Transformerscited by 6,968
  6. 2023SDXL: Improving Latent Diffusion Models for High-Resolution Image Synthesiscited by 4,356
  7. 2024Scaling Rectified Flow Transformers for High-Resolution Image Synthesiscited by 4,181
  8. 2022LAION-5B: An open large-scale dataset for training next generation image-text modelscited by 4,152
  9. 2018UMAP: Uniform Manifold Approximation and Projection for Dimension Reductioncited by 2,898
  10. 2020Supervised Contrastive Learningcited by 2,829
  11. 2020The Many Faces of Robustness: A Critical Analysis of Out-of-Distribution Generalizationcited by 1,850
  12. 2024Depth Anything V2cited by 1,422
36 more
  1. 2023Kosmos-2: Grounding Multimodal Large Language Models to the Worldcited by 1,089
  2. 2022Diffusion Models in Vision: A Surveycited by 1,015
  3. 2024Grounded SAM: Assembling Open-World Models for Diverse Visual Taskscited by 929
  4. 2019CNN-generated images are surprisingly easy to spot... for nowcited by 847
  5. 2021CrossViT: Cross-Attention Multi-Scale Vision Transformer for Image Classificationcited by 806
  6. 2019Multimodal Transformer for Unaligned Multimodal Language Sequencescited by 645
  7. 2017Generalised Dice overlap as a deep learning loss function for highly unbalanced segmentationscited by 620
  8. 2021DRAEM -- A discriminatively trained reconstruction embedding for surface anomaly detectioncited by 514
  9. 2020Leveraging Frequency Analysis for Deep Fake Image Recognitioncited by 504
  10. 2023Towards Universal Fake Image Detectors that Generalize Across Generative Modelscited by 490
  11. 2023Inpaint Anything: Segment Anything Meets Image Inpaintingcited by 233
  12. 2024A Sanity Check for AI-generated Image Detectioncited by 189
  13. 2024PIXART-δ: Fast and Controllable Image Generation with Latent Consistency Modelscited by 141
  14. 2023Hierarchical Fine-Grained Image Forgery Detection and Localizationcited by 134
  15. 2022TruFor: Leveraging all-round clues for trustworthy image forgery detection and localizationcited by 130
  16. 2024C2P-CLIP: Injecting Category Common Prompt in CLIP to Enhance Generalization in Deepfake Detectioncited by 113
  17. 2024FakeShield: Explainable Image Forgery Detection and Localization via Multi-modal Large Language Modelscited by 113
  18. 2024SIDA: Social Media Image Deepfake Detection, Localization and Explanation with Large Multimodal Modelcited by 108
  19. 2022ObjectFormer for Image Manipulation Detection and Localizationcited by 78
  20. 2020The Edge of Depth: Explicit Constraints between Segmentation and Depthcited by 72
  21. 2025Spot the Fake: Large Multimodal Model-Based Synthetic Image Detection with Artifact Explanationcited by 68
  22. 2020Siamese Network for RGB-D Salient Object Detection and Beyondcited by 67
  23. 2023Shadows Don't Lie and Lines Can't Bend! Generative Models don't know Projective Geometry...for nowcited by 61
  24. 2023Online Detection of AI-Generated Imagescited by 55
  25. 2024Community Forensics: Using Thousands of Generators to Train Fake Image Detectorscited by 50
  26. 2024FIRE: Robust Detection of Diffusion-Generated Images via Frequency-Guided Reconstruction Errorcited by 47
  27. 2023AutoSplice: A Text-prompt Manipulated Image Dataset for Media Forensicscited by 46
  28. 2024Structure Matters: Tackling the Semantic Discrepancy in Diffusion Models for Image Inpaintingcited by 40
  29. 2024Mesoscopic Insights: Orchestrating Multi-scale & Hybrid Architecture for Image Manipulation Localizationcited by 35
  30. 2024WildFake: A Large-scale Challenging Dataset for AI-Generated Images Detectioncited by 31
  31. 2024Frequency Masking for Universal Deepfake Detectioncited by 23
  32. 2025OpenSDI: Spotting Diffusion-Generated Images in the Open Worldcited by 23
  33. 2024GIM: A Million-scale Benchmark for Generative Image Manipulation Detection and Localizationcited by 19
  34. 2024SAFIRE: Segment Any Forged Image Regioncited by 18
  35. 2024On Pretraining Data Diversity for Self-Supervised Learningcited by 7
  36. 2024Adaptive Deviation Learning for Visual Anomaly Detection with Data Contaminationcited by 4

Built on by

Cited by year (2026 to 14 Oct)

Citing papers per year: 2025: 1; 2026: 3.

4 papers.

  1. 2026EvoGuard: An Extensible Agentic RL-based Framework for Practical and Evolving AI-Generated Image Detectioncited by 6
  2. 2025Revisiting Image Manipulation Localization under Realistic Manipulation Scenarioscited by 3
  3. 2026Dissecting Agentic Forensics: The Role of Triage, Prompting, and Evidence Arbitration in Open-World Fake Image Detectioncited by 0
  4. 2026Less Supervision, Better Generalization: Weakly Supervised Fake Region Localization in Diffusion-Edited Imagescited by 0

Similar, not cited

Close in wording, no citation either way

  1. 2023Differential Diffusion: Giving Each Pixel Its Strength
  2. 2025DEAL-300K: Diffusion-based Editing Area Localization with a 300K-Scale Dataset and Frequency-Prompted Baseline
  3. 2023Diffusion Brush: A Latent Diffusion Model-based Editing Tool for AI-generated Images
  4. 2025X-Edit: Detecting and Localizing Edits in Images Altered by Text-Guided Diffusion Models
  5. 2025End4: End-to-end Denoising Diffusion for Diffusion-Based Inpainting Detection
  6. 2023Diffusion Handles: Enabling 3D Edits for Diffusion Models by Lifting Activations to 3D
  7. 2023Towards Coherent Image Inpainting Using Denoising Diffusion Implicit Models
  8. 2025Edit2Perceive: Image Editing Diffusion Models Are Strong Dense Perceivers