Research map / Image generation / Text-to-image diffusion
Few-shot Defect Image Generation based on Consistency Modeling
Qingfeng Shi, Jing Wei, Fei Shen, Zhengtao Zhang
Read the paperCheck your idea against 137,217 papers
AI-extracted. The measured effects below were read by AI from the paper's full text (no summary yet). Numbers and claims can be wrong or missing context, so check the paper before you cite or build on it.
Measured effects reported in the paper
| What was changed | Where | Measured | Effect |
|---|---|---|---|
| add the fusion part to the generation module (ablation) | DefectDiffu · MVTecAD average | FID: 149.83 -> 69.35 | Improved |
| add adaptive attention-enhanced loss (AAL) (ablation) | DefectDiffu · MVTecAD average | FID: 152.77 -> 69.35 | Improved |
| use multiple single-object text prompts instead of one fused prompt (ablation) | DefectDiffu · MVTecAD average | FID: 197.55 -> 69.35 | Improved |
| generate at multiple defect strength scales w_d instead of single scale (ablation) | DefectDiffu · MVTecAD average | LPIPS: 0.20 -> 0.21 | Improved |
| augment segmentation training with DefectDiffu synthetic image-mask pairs | UNet · MVTecAD average | mIoU: 60.3 -> 67.1 | Improved |
| augment segmentation training with DFMGAN synthetic data | UNet · MVTecAD average | mIoU: 60.3 -> 60.3 | No clear effect |
| augment segmentation training with Defect-Gen synthetic data | UNet · MVTecAD average | mIoU: 60.3 -> 63.5 | Improved |
| use DefectDiffu instead of Defect-Gen for diversity | DefectDiffu · MVTecAD average | LPIPS: 0.24 -> 0.21 | Hurt |
Closest papers by meaning
- DefectFill: Realistic Defect Generation with Inpainting Diffusion Model for Visual Inspection (CVPR 2025 · similarity 0.90)
- Wavelet and Prototype Augmented Query-based Transformer for Pixel-level Surface Defect Detection (CVPR 2025 · similarity 0.85)
- Dual-Interrelated Diffusion Model for Few-Shot Anomaly Image Generation (CVPR 2025 · similarity 0.84)
- Automated Defect Report Generation for Enhanced Industrial Quality Control (AAAI 2024 · similarity 0.84)
- Exploring the Relationship Between Samples and Masks for Robust Defect Localization (AAAI 2025 · similarity 0.84)
- MVREC: A General Few-shot Defect Classification Model Using Multi-View Region-Context (AAAI 2025 · similarity 0.84)
Writing a paper in this area? scime draft is a free LaTeX editor with AI included, real-time co-authors and a check of your references and venue rules before you submit. It also imports OpenAI Prism projects.
