mlsanityData to 14 Oct 2026

Training-Free Anomaly Generation via Dual-Attention Enhancement in Diffusion Model

arXiv 2508.11550 ↗First posted 15 Aug 2025cs.CVcited by 0

On the map: Diffusion models → Flow matching, time series and science

Abstract (arXiv)

Industrial anomaly detection (AD) plays a significant role in manufacturing where a long-standing challenge is data scarcity. A growing body of works have emerged to address insufficient anomaly data via anomaly generation. However, these anomaly generation methods suffer from lack of fidelity or need to be trained with extra data. To this end, we propose a training-free anomaly generation framework dubbed AAG, which is based on Stable Diffusion (SD)'s strong generation ability for effective anomaly image generation. Given a normal image, mask and a simple text prompt, AAG can generate realistic and natural anomalies in the specific regions and simultaneously keep contents in other regions unchanged. In particular, we propose Cross-Attention Enhancement (CAE) to re-engineer the cross-attention mechanism within Stable Diffusion based on the given mask. CAE increases the similarity between visual tokens in specific regions and text embeddings, which guides these generated visual tokens in accordance with the text description. Besides, generated anomalies need to be more natural and plausible with object in given image. We propose Self-Attention Enhancement (SAE) which improves similarity between each normal visual token and anomaly visual tokens. SAE ensures that generated anomalies are coherent with original pattern. Extensive experiments on MVTec AD and VisA datasets demonstrate effectiveness of AAG in anomaly generation and its utility. Furthermore, anomaly images generated by AAG can bolster performance of various downstream anomaly inspection tasks.

Key facts. No citing papers in the map yet (data through 14 Oct 2026). Placed in Flow matching, time series and science, one of 14 directions in Diffusion models. Builds on 23 of its 31 references that are in the map. First posted on arXiv 15 Aug 2025.

Builds on

23 of 31 references.

Closest ancestors · 1

Same direction, newest first

  1. 2023AnomalyDiffusion: Few-Shot Anomaly Image Generation with Diffusion Model10 Dec

Other references · 22

  1. 2020Denoising Diffusion Probabilistic Modelscited by 21,076
  2. 2015U-Net: Convolutional Networks for Biomedical Image Segmentationcited by 19,713
  3. 2021Diffusion Models Beat GANs on Image Synthesiscited by 8,255
  4. 2015Deep Unsupervised Learning using Nonequilibrium Thermodynamicscited by 7,058
  5. 2022Classifier-Free Diffusion Guidancecited by 5,847
  6. 2023SDXL: Improving Latent Diffusion Models for High-Resolution Image Synthesiscited by 4,356
  7. 2016Improved Techniques for Training GANscited by 2,970
  8. 2022Prompt-to-Prompt Image Editing with Cross Attention Controlcited by 2,431
  9. 2020PaDiM: a Patch Distribution Modeling Framework for Anomaly Detection and Localizationcited by 594
  10. 2021CutPaste: Self-Supervised Learning for Anomaly Detection and Localizationcited by 531
  11. 2021DRAEM -- A discriminatively trained reconstruction embedding for surface anomaly detectioncited by 514
  12. 2021Few-shot Image Generation via Cross-domain Correspondencecited by 207
10 more
  1. 2022Improving Sample Quality of Diffusion Models Using Self-Attention Guidancecited by 154
  2. 2024Towards Understanding Cross and Self-Attention in Stable Diffusion for Text-Guided Image Editingcited by 135
  3. 2024RealNet: A Feature Selection Network with Realistic Synthetic Anomaly for Anomaly Detectioncited by 116
  4. 2021Defect-GAN: High-Fidelity Defect Synthesis for Automated Defect Inspectioncited by 77
  5. 2022MemSeg: A semi-supervised method for image surface defect detection using differences and commonalitiescited by 68
  6. 2023Few-Shot Defect Image Generation via Defect-Aware Feature Manipulationcited by 67
  7. 2024Attentive Eraser: Unleashing Diffusion Model's Object Removal Potential via Self-Attention Redirection Guidancecited by 63
  8. 2023Collaborative Discrepancy Optimization for Reliable Image Anomaly Localizationcited by 57
  9. 2024AnomalyXFusion: Multi-modal Anomaly Synthesis with Diffusioncited by 21
  10. 2024AnomalyControl: Learning Cross-modal Semantic Features for Controllable Anomaly Synthesiscited by 7

Built on by

Not cited yet.

Similar, not cited

Close in wording, no citation either way

  1. 2024Unseen Visual Anomaly Generation
  2. 2024Dual-Interrelated Diffusion Model for Few-Shot Anomaly Image Generation
  3. 2024AnomalySD: Few-Shot Multi-Class Anomaly Detection with Stable Diffusion Model
  4. 2024A Novel Approach to Industrial Defect Generation through Blended Latent Diffusion Model with Online Adaptation
  5. 2025Generate Aligned Anomaly: Region-Guided Few-Shot Anomaly Image-Mask Pair Synthesis for Industrial Inspection
  6. 2026FLASH: A "Generate Once, Synthesize Many" Framework for Synthetic Anomaly Generation in Industrial Anomaly Detection
  7. 2025Few-Shot Anomaly-Driven Generation for Anomaly Classification and Segmentation
  8. 2025Anomaly Detection by Effectively Leveraging Synthetic Images