A curated list of resources including papers, datasets, and relevant links pertaining to few-shot defect image generation.Few-shot defect image generation aims to synthesize diverse and realistic defect images using only a limited number of annotated anomaly samples. It primarily focuses on augmenting data for industrial scenarios to improve downstream tasks such as anomaly detection, localization, and segmentation, despite severe data scarcity.
Contributions are welcome. If you wish to contribute, feel free to send a pull request. If you have suggestions for new sections to be included, please raise an issue and discuss before sending a pull request.
- Yunkang Cao, Xiaohao Xu, Jiangning Zhang, Yuqi Cheng, Xiaonan Huang, Guansong Pang, Weiming Shen: "A Survey on Visual Anomaly Detection: Challenge, Approach, and Prospect." arXiv (2024) [pdf]
- Yang Liu, Jing Liu, Chengfang Li, Rui Xi, Wenchao Li, Liang Cao, Jin Wang, Laurence T. Yang, Junsong Yuan, Wei Zhou: "Anomaly Detection and Generation with Diffusion Models: A Survey." arXiv (2025) [pdf]
- Xichen Xu, Yanshu Wang, Yawen Huang, Jiaqi Liu, Xiaoning Lei, Guoyang Xie, Guannan Jiang, Zhichao Lu: "A Survey on Industrial Anomalies Synthesis." arXiv (2025) [pdf]
- Dongyun Lin, Yanpeng Cao, Wenbing Zhu, Yiqun Li: "Few-Shot Defect Segmentation Leveraging Abundant Normal Training Samples Through Normal Background Regularization and Crop-and-Paste Operation." ICME (2021) [pdf]
- Vitjan Zavrtanik, Matej Kristan, Danijel Skočaj: "DRAEM -- A discriminatively trained reconstruction embedding for surface anomaly detection." ICCV (2021) [pdf] [code]
- Chun-Liang Li, Kihyuk Sohn, Jinsung Yoon, Tomas Pfister: "CutPaste: Self-Supervised Learning for Anomaly Detection and Localization." CVPR (2021) [pdf] [code]
- Hui Zhang, Zuxuan Wu, Zheng Wang, Zhineng Chen, Yu-Gang Jiang: "Prototypical Residual Networks for Anomaly Detection and Localization." CVPR (2023) [pdf] [code]
- Shuanlong Niu, Bin Li, Xinggang Wang, HuiLin: "Defect Image Sample Generation With GAN for Improving Defect Recognition." IEEE TASE (2020) [pdf]
- Gongjie Zhang, Kaiwen Cui, Tzu-Yi Hung, Shijian Lu: "Defect-GAN: High-Fidelity Defect Synthesis for Automated Defect Inspection." WACV (2021) [pdf]
- Yuxuan Duan, Yan Hong, Li Niu, Liqing Zhang: "Few-Shot Defect Image Generation via Defect-Aware Feature Manipulation." AAAI (2023) [pdf] [code]
- Musawar Ali, Nicola Fioraio, Samuele Salti, Luigi Di Stefano: "AnomalyControl: Few-Shot Anomaly Generation by ControlNet Inpainting.*" IEEE Access (2024) [pdf]
- Qianzi Yu, Kai Zhu, Yang Cao, Feijie Xia, Yu Kang: "TF²: Few-Shot Text-Free Training-Free Defect Image Generation for Industrial Anomaly Inspection." IEEE TCSVT (2024) [pdf]
- Adnan Md Tayeb, Hope Leticia Nakayiza , Heejae Shin, Seungmin Lee, Chaesoo Lee, YeongHun Lee, Dong-Seong Kim, Jae-Min Lee: "DefectGen: Few-Shot Defect Image Generation Using Stable Diffusion for Steel Surface Analysis." IEEE ICTC (2024) [pdf]
- Guan Gui, Bin-Bin Gao, Jun Liu, Chengjie Wang, Yunsheng Wu: "Few-Shot Anomaly-Driven Generation for Anomaly Classification and Segmentation." ECCV (2024) [pdf] [code]
- Qingfeng Shi, Jing Wei, Fei Shen, Zhengtao Zhang: "Few-shot Defect Image Generation based on Consistency Modeling." ECCV (2024) [pdf] [code]
- Shuai Yang, Zhifei Chen, Pengguang Chen, Xi Fang, Yixun Liang, Shu Liu, Yingcong Chen: "Defect Spectrum: A Granular Look of Large-Scale Defect Datasets with Rich Semantics." ECCV (2024) [pdf] [code]
- Teng Hu, Jiangning Zhang, Ran Yi, Yuzhen Du, Xu Chen, Liang Liu, Yabiao Wang, Chengjie Wang: "AnomalyDiffusion: Few-Shot Anomaly Image Generation with Diffusion Model." AAAI (2024) [pdf] [code]
- Ximiao Zhang, Min Xu, Xiuzhuang Zhou: "RealNet: A Feature Selection Network with Realistic Synthetic Anomaly for Anomaly Detection." CVPR (2024) [pdf] [code]
- Zhewei Dai, Shilei Zeng, Haotian Liu, Xurui Li, Feng Xue, Yu Zhou: "SeaS: Few-shot Industrial Anomaly Image Generation with Separation and Sharing Fine-tuning." ICCV (2025) [pdf] [code]
- Ying Jin, Jinlong Peng, Qingdong He, Teng Hu, Jiafu Wu, Hao Chen, Haoxuan Wang, Wenbing Zhu, Mingmin Chi, Jun Liu, Yabiao Wang: "Dual-Interrelated Diffusion Model for Few-Shot Anomaly Image Generation." CVPR (2025) [pdf] [code]
- Han Sun, Yunkang Cao, Hao Dong, Olga Fink: "Unseen Visual Anomaly Generation." CVPR (2025) [pdf] [code]
- Ruyi Xu, Yen-Tzu Chiu, Tai-I Chen, Oscar Chew, Yung-Yu Chuang, Wen-Huang Cheng: "Training-Free Industrial Defect Generation with Diffusion Models." ICCV (2025) [pdf] [code]
- Hannah M. Schlüter, Jeremy Tan, Benjamin Hou, Bernhard Kainz: "Natural Synthetic Anomalies for Self-Supervised Anomaly Detection and Localization." ECCV (2022) [pdf] [code]
- MVTec AD: 5354 high-resolution RGB images across 15 categories — 10 object classes (e.g., bottle, cable, metal nut) and 5 texture classes (e.g., carpet, leather, tile). Each category includes a variety of anomaly types such as scratches, dents, contaminations, or missing parts, totaling 73 distinct defect types. [link]
- VisA: 12 subsets with 10,821 images (9,621 normal, 1,200 anomalous) covering diverse objects. Anomalies involve surface defects like scratches and dents, and structural defects such as misplacement or missing parts. [link]
- MVTec 3D-AD: over 4,000 high-resolution 3D scans across 10 object categories for unsupervised 3D anomaly detection and localization, with defect-free training data, defective test samples, and precise ground-truth annotations. [link]
- Kaputt: a large-scale and realistic dataset for visual defect detection, containing 238K images, 29K defective samples, and 48K unique items with detailed annotations. It features diverse appearances, poses, and real-world noise, offering a challenging benchmark that exposes the limitations of current defect detection methods. [link]