[1]蒋世杰,夏秀山,翟伟,等.基于ODE扩散模型的多类异常检测和定位[J].智能系统学报,2025,20(2):376-388.[doi:10.11992/tis.202402022]
JIANG Shijie,XIA Xiushan,ZHAI Wei,et al.ODE diffusion model for multiclass anomaly detection and localization[J].CAAI Transactions on Intelligent Systems,2025,20(2):376-388.[doi:10.11992/tis.202402022]
点击复制
JIANG Shijie,XIA Xiushan,ZHAI Wei,et al.ODE diffusion model for multiclass anomaly detection and localization[J].CAAI Transactions on Intelligent Systems,2025,20(2):376-388.[doi:10.11992/tis.202402022]
基于ODE扩散模型的多类异常检测和定位
《智能系统学报》[ISSN 1673-4785/CN 23-1538/TP] 卷:
20
期数:
2025年第2期
页码:
376-388
栏目:
学术论文—机器感知与模式识别
出版日期:
2025-03-05
- Title:
- ODE diffusion model for multiclass anomaly detection and localization
- Keywords:
- defect detection; anomaly detection; anomaly localization; diffusion model; denoising network; ordinary differential equations; unsupervised learning; timestep-perceptive
- 分类号:
- TP391
- 摘要:
- 多类异常检测和定位旨在训练一个单一模型，在多类场景下该模型能够识别出偏离正常的异常区域。最近基于扩散模型的方法在该项任务中表现出色而开始受到关注，然而，现有方法侧重于改进扩散模型去噪网络，通过添加更多约束，保持多步生成的高一致性，实现更高的重构性能，但更多的采样步数也意味着更高的计算开销。为此，本文提出了一种基于常微分方程(ordinary differential equations, ODE)扩散模型的多类异常检测和定位方法，只需一步即可实现高质量的重构生成，同时引入时间步感知网络来缓解采样步数少可能导致的一致性和恒等捷径问题，从而进一步提高重构质量。在通用的基准数据集MVTec-AD上进行的实验结果表明，本文方法在精度上可与当前最先进方法相媲美，但是计算量更低速度更快，满足了工业异常检测和定位的高精度和实时性需求。
- Abstract:
- Multiclass anomaly detection and localization methods aim to train a single model capable of identifying anomalous regions that deviate from normal across multiple categories. Diffusion-based methods have recently attracted attention due to their excellent performance in this task. However, existing methods concentrate on enhancing the denoising network of the diffusion model by adding more constraints to ensure high consistency in multistep generation and achieve superior reconstruction performance. Additional sampling steps also lead to higher computational costs. We propose a novel multiclass anomaly detection and localization method called TimeNet to address these issues. It is based on the diffusion model of ordinary differential equations and achieves high-quality reconstruction with only one-step generation. We introduce a time-perceptive network to address the consistency and identity shortcut problems that may arise from small sampling steps, which further improves the reconstruction quality. Experiments on the most popular benchmark MVTec-AD dataset demonstrate that our TimeNet competes with the current state-of-the-art methods in terms of accuracy while requiring less computational effort and achieving faster speeds. The high accuracy and real-time performance of TimeNet satisfy the requirements for industrial anomaly detection and localization.
- 参考文献/References:
-
[1] BERGMANN P, FAUSER M, SATTLEGGER D, et al. MVTec AD—a comprehensive real-world dataset for unsupervised anomaly detection[C]//2019 IEEE/CVF Conference on Computer Vision and Pattern Recognition. Long Beach: IEEE, 2019: 9584-9592.
[2] 吕承侃, 沈飞, 张正涛, 等. 图像异常检测研究现状综述[J]. 自动化学报, 2022, 48(6): 1402-1428.
LYU Chengkan, SHEN Fei, ZHANG Zhengtao, et al. Review of image anomaly detection[J]. Acta automatica sinica, 2022, 48(6): 1402-1428.
[3] 陶显, 侯伟, 徐德. 基于深度学习的表面缺陷检测方法综述[J]. 自动化学报, 2021, 47(5): 1017-1034.
TAO Xian, HOU Wei, XU De. A survey of surface defect detection methods based on deep learning[J]. Acta automatica sinica, 2021, 47(5): 1017-1034.
[4] 伍麟, 郝鸿宇, 宋友. 基于计算机视觉的工业金属表面缺陷检测综述[J]. 自动化学报, 2024, 50(7): 1261-1283.
WU Lin, HAO Hongyu, SONG You. A review of metal surface defect detection based on computer vision[J]. Acta automatica sinica, 2024, 50(7): 1261-1283.
[5] 黄健, 郑春厚, 章军, 等. 基于小样本度量迁移学习的表面缺陷检测[J]. 模式识别与人工智能, 2021, 34(5): 407-414.
HUANG Jian, ZHENG Chunhou, ZHANG Jun, et al. Few-shot metric transfer learning network for surface defect detection[J]. Pattern recognition and artificial intelligence, 2021, 34(5): 407-414.
[6] 邢鹏, 蒋鑫, 潘永华, 等. 基于特征约束蒸馏学习的视觉异常检测[J]. 软件学报, 2023, 34(9): 4378-4391.
XING Peng, JIANG Xin, PAN Yonghua, et al. Feature constrained restricted distillation learning for visual anomaly detection[J]. Journal of software, 2023, 34(9): 4378-4391.
[7] 王延春秋, 葛泉波, 刘华平. 石板材表面缺陷检测的无监督学习方法[J]. 智能系统学报, 2023, 18(6): 1344-1351.
WANG Yanchunqiu, GE Quanbo, LIU Huaping. Unsupervised learning method for surface defect detection of slate materials[J]. CAAI transactions on intelligent systems, 2023, 18(6): 1344-1351.
[8] 孙博言, 王洪元, 刘乾, 等. 基于多尺度和注意力机制的混合监督金属表面缺陷检测[J]. 智能系统学报, 2023, 18(4): 886-893.
SUN Boyan, WANG Hongyuan, LIU Qian, et al. Hybrid supervised metal surface defect detection based on multi-scale and attention[J]. CAAI transactions on intelligent systems, 2023, 18(4): 886-893.
[9] FERNANDO T, GAMMULLE H, DENMAN S, et al. Deep learning for medical anomaly detection—a survey[J]. ACM computing surveys, 2022, 54(7): 1-37.
[10] 胡显耀, 靳聪明. 基于扩散常微分方程的医学图像异常检测[J]. 浙江理工大学学报(自然科学版), 2024, 49(6): 851-860.
HU Xianyao, JIN Congming. Medical image anomaly detection based on diffusion ordinary differential equations[J]. Journal of Zhejiang Sci-Tech University (natural sciences edition), 2024, 49(6): 851-860.
[11] NAYAK R, PATI U C, DAS S K. A comprehensive review on deep learning-based methods for video anomaly detection[J]. Image and vision computing, 2021, 106: 104078.
[12] 胡正平, 赵梦瑶, 辛丙一. 结合全局与局部视频表示的视频异常检测算法[J]. 模式识别与人工智能, 2020, 33(2): 133-140.
HU Zhengping, ZHAO Mengyao, XIN Bingyi. Video anomaly detection algorithm combining global and local video representation[J]. Pattern recognition and artificial intelligence, 2020, 33(2): 133-140.
[13] 黄超, 胡志军, 徐勇, 等. 基于视觉的车辆异常行为检测综述[J]. 模式识别与人工智能, 2020, 33(3): 234-248.
HUANG Chao, HU Zhijun, XU Yong, et al. Vision-based abnormal vehicle behavior detection: a survey[J]. Pattern recognition and artificial intelligence, 2020, 33(3): 234-248.
[14] ZAVRTANIK V, KRISTAN M, SKO?AJ D. DR?M–A discriminatively trained reconstruction embedding for surface anomaly detection[C]//2021 IEEE/CVF International Conference on Computer Vision. Montreal: IEEE, 2021: 8310-8319.
[15] GUDOVSKIY D, ISHIZAKA S, KOZUKA K. CFLOW-AD: real-time unsupervised anomaly detection with localization via conditional normalizing flows[C]//2022 IEEE/CVF Winter Conference on Applications of Computer Vision. Waikoloa: IEEE, 2022: 1819-1828.
[16] ROTH K, PEMULA L, ZEPEDA J, et al. Towards total recall in industrial anomaly detection[C]//2022 IEEE/CVF Conference on Computer Vision and Pattern Recognition. New Orleans: IEEE, 2022: 14298-14308.
[17] YOU Zhiyuan, CUI Lei, SHEN Yujun, et al. A unified model for multi-class anomaly detection[C]//Proceedings of the 36th International Conference on Neural Information Processing Systems. New Orleans: Curran Associates Inc., 2022: 4571-4584.
[18] ZHAO Ying. OmniAL: a unified CNN framework for unsupervised anomaly localization[C]//2023 IEEE/CVF Conference on Computer Vision and Pattern Recognition. Vancouver: IEEE, 2023: 3924-3933.
[19] LU Ruiying, WU Yujie, TIAN Long, et al. Hierarchical vector quantized transformer for multi-class unsupervised anomaly detection[C]//Proceedings of the 37th International Conference on Neural Information Processing Systems. New Orleans: Curran Associates Inc., 2023: 8487-8500.
[20] LU Fanbin, YAO Xufeng, FU C W, et al. Removing anomalies as noises for industrial defect localization[C]//2023 IEEE/CVF International Conference on Computer Vision. Paris: IEEE, 2023: 16120-16129.
[21] HE Haoyang, ZHANG Jiangning, CHEN Hongxu, et al. DiAD: a diffusion-based framework for multi-class anomaly detection[EB/OL]. (2023-12-11)[2024-02-22]. https://arxiv.org/abs/2312.06607.
[22] WANG Guodong, HAN Shumin, DING Errui, et al. Student-teacher feature pyramid matching for anomaly detection[EB/OL]. (2021-03-07)[2024-02-22]. https://arxiv.org/abs/2103.04257.
[23] DENG Hanqiu, LI Xingyu. Anomaly detection via reverse distillation from one-class embedding[C]//2022 IEEE/CVF Conference on Computer Vision and Pattern Recognition. New Orleans: IEEE, 2022: 9727-9736.
[24] TIEN T D, NGUYEN A T, TRAN N H, et al. Revisiting reverse distillation for anomaly detection[C]//2023 IEEE/CVF Conference on Computer Vision and Pattern Recognition. Vancouver: IEEE, 2023: 24511-24520.
[25] RUDOLPH M, WANDT B, ROSENHAHN B. Same same but DifferNet: semi-supervised defect detection with normalizing flows[C]//2021 IEEE Winter Conference on Applications of Computer Vision. Waikoloa: IEEE, 2021: 1906-1915.
[26] LEI Jiarui, HU Xiaobo, WANG Yue, et al. PyramidFlow: high-resolution defect contrastive localization using pyramid normalizing flow[C]//2023 IEEE/CVF Conference on Computer Vision and Pattern Recognition. Vancouver: IEEE, 2023: 14143-14152.
[27] 毛国君, 吴星臻, 邢树礼. 基于多尺度流模型的视觉异常检测研究[J]. 自动化学报, 2024, 50(3): 640-648.
MAO Guojun, WU Xingzhen, XING Shuli. Research on visual anomaly detection based on multi-scale normalizing flow[J]. Acta automatica sinica, 2024, 50(3): 640-648.
[28] GONG Dong, LIU Lingqiao, LE V, et al. Memorizing normality to detect anomaly: memory-augmented deep autoencoder for unsupervised anomaly detection[C]//2019 IEEE/CVF International Conference on Computer Vision. Seoul: IEEE, 2019: 1705-1714.
[29] AKCAY S, ATAPOUR-ABARGHOUEI A, BRECKON T P. GANomaly: semi-supervised anomaly detection via adversarial training[C]//14th Asian Conference on Computer Vision. Perth: Springer International Publishing, 2019: 622-637.
[30] PERERA P, NALLAPATI R, XIANG Bing. OCGAN: one-class novelty detection using GANs with constrained latent representations[C]//2019 IEEE/CVF Conference on Computer Vision and Pattern Recognition. Long Beach: IEEE, 2019: 2893-2901.
[31] WYATT J, LEACH A, SCHMON S M, et al. AnoDDPM: anomaly detection with denoising diffusion probabilistic models using simplex noise[C]//2022 IEEE/CVF Conference on Computer Vision and Pattern Recognition Workshops. New Orleans: IEEE, 2022: 649-655.
[32] ZHANG Xinyi, LI Naiqi, LI Jiawei, et al. Unsupervised surface anomaly detection with diffusion probabilistic model[C]//2023 IEEE/CVF International Conference on Computer Vision. Paris: IEEE, 2023: 6759-6768.
[33] MOUSAKHAN A, BROX T, TAYYUB J. Anomaly detection with conditioned denoising diffusion models[EB/OL]. (2023-05-25)[2024-02-22]. https://arxiv.org/abs/2305.15956.
[34] DHARIWAL P, NICHOL A. Diffusion models beat gans on image synthesis[C]//Proceedings of the 35th International Conference on Neural Information Processing Systems. Vancouver: Curran Associates Inc., 2021, 34: 8780-8794.
[35] SONG Jiaming, MENG Chenlin, Ermon S. Denoising diffusion implicit models[C]//International Conference on Learning Representations. Vienna: OpenReview.net, 2021: 1-20.
[36] NICHOL A, DHARIWAL P. Improved denoising diffusion probabilistic models[C]//International Conference on Machine Learning. Virtual: PMLR, 2021: 8162-8171.
[37] LIU Xingchao, GONG Chengyue, LIU Qiang. Flow straight and fast: learning to generate and transfer data with rectified flow[C]//International Conference on Learning Representations. Kigali: OpenReview. net, 2023: 1-33.
[38] LIU Xingchao, ZHANG Xiwen, MA Jianzhu, et al. InstaFlow: one step is enough for high-quality diffusion-based text-to-image generation[EB/OL]. (2023-09-12)[2024-02-22]. https://arxiv.org/abs/2309.06380.
[39] RONNEBERGER O, FISCHER P, BROX T. U-net: convolutional networks for biomedical image segmentation[C]//Medical Image Computing and Computer-Assisted Intervention. Munich: Springer International Publishing, 2015: 234-241.
[40] SOHL-DICKSTEIN J, WEISS E A, MAHESWARANATHAN N, et al. Deep unsupervised learning using nonequilibrium thermodynamics[C]//International Conference on Machine Learning. Lille: PMLR, 2015: 2256-2265.
[41] HO J, JAIN A, ABBEEL P. Denoising diffusion probabilistic models[C]//Proceedings of the 34th International Conference on Neural Information Processing Systems. Virtual: Curran Associates Inc., 2020, 33: 6840-6851.
[42] ROMBACH R, BLATTMANN A, LORENZ D, et al. High-resolution image synthesis with latent diffusion models[C]//2022 IEEE/CVF Conference on Computer Vision and Pattern Recognition. New Orleans: IEEE, 2022: 10674-10685.
[43] ZHANG Lyumin, RAO Anyi, AGRAWALA M. Adding conditional control to text-to-image diffusion models[C]//2023 IEEE/CVF International Conference on Computer Vision. Paris: IEEE, 2023: 3813-3824.
[44] LU Cheng, ZHOU Yuhao, BAO Fan, et al. Dpm-solver: a fast ode solver for diffusion probabilistic model sampling in around 10 steps[C]//Proceedings of the 36th International Conference on Neural Information Processing Systems. New Orleans: Curran Associates Inc., 2022, 35: 5775-5787.
[45] ZHOU Zhenyu, CHEN Defang, WANG Can, et al. Fast ODE-based sampling for diffusion models in around 5 steps[EB/OL]. (2023-11-30)[2024-02-22]. https://arxiv.org/abs/2312.00094.
[46] SONG Yang, DHARIWAL P, CHEN M, et al. Consistency models[EB/OL]. (2023-03-02)[2024-02-22]. https://arxiv.org/abs/2303.01469.
[47] YIN Tianwei, GHARBI M, ZHANG R, et al. One-step diffusion with distribution matching distillation[EB/OL]. (2023-11-30)[2024-02-22]. https://arxiv.org/abs/2311.18828.
[48] SONG Yang, SOHL-DICKSTEIN J, KINGMA D P, et al. Score-based generative modeling through stochastic differential equations[C]//International Conference on Learning Representations. Vienna: OpenReview.net, 2021: 1-36.
[49] CIMPOI M, MAJI S, KOKKINOS I, et al. Describing textures in the wild[C]//2014 IEEE Conference on Computer Vision and Pattern Recognition. Columbus: IEEE, 2014: 3606-3613.
[50] ZHANG R, ISOLA P, EFROS A A, et al. The unreasonable effectiveness of deep features as a perceptual metric[C]//2018 IEEE/CVF Conference on Computer Vision and Pattern Recognition. Salt Lake City: IEEE, 2018: 586-595.
[51] HE Kaiming, ZHANG Xiangyu, REN Shaoqing, et al. Deep residual learning for image recognition[C]//2016 IEEE Conference on Computer Vision and Pattern Recognition. Las Vegas: IEEE, 2016: 770-778.
[52] ZAGORUYKO S, KOMODAKIS N. Wide residual networks[EB/OL]. (2016-05-03)[2024-02-22]. https://arxiv.org/abs/1605.07146.
[53] SIMONYAN K, ZISSERMAN A. Very deep convolutional networks for large-scale image recognition[C]//International Conference on Learning Representations. San Diego: OpenReview.net, 2015: 1–14.
[54] TAN Mingxing, LE Q. EfficientNet: rethinking model scaling for convolutional neural networks[C]//Proceedings of the 36th International Conference on Machine Learning. Los Angeles: PMLR, 2019: 6105–6114.
- 相似文献/References:
-
[1]田国会,吉艳青,李晓磊.家庭智能空间下基于场景的人的行为理解[J].智能系统学报,2010,5(1):57.
TIAN Guo-hui,JI Yan-qing,LI Xiao-lei.Human behaviors understanding based on scene knowledge in home intelligent space[J].CAAI Transactions on Intelligent Systems,2010,5():57.
[2]吴家伟,严京旗,方志宏,等.基于图像显著性特征的铸坯表面缺陷检测[J].智能系统学报,2012,7(1):75.
WU Jiawei,YAN Jingqi,FANG Zhihong,et al.Defect detection on a steel slab surface based on the characteristics of an image’s saliency region[J].CAAI Transactions on Intelligent Systems,2012,7():75.
[3]赵冠哲,齐建鹏,于彦伟,等.移动社交网络异常签到在线检测算法[J].智能系统学报,2017,12(5):752.[doi:10.11992/tis.201706027]
ZHAO Guanzhe,QI Jianpeng,YU Yanwei,et al.Online check-in outlier detection method in mobile social networks[J].CAAI Transactions on Intelligent Systems,2017,12():752.[doi:10.11992/tis.201706027]
[4]张玉玲,尹传环.依特征频率的安卓恶意软件异常检测的研究[J].智能系统学报,2018,13(2):168.[doi:10.11992/tis.201609016]
ZHANG Yuling,YIN Chuanhuan.Android malware outlier detection based on feature frequency[J].CAAI Transactions on Intelligent Systems,2018,13():168.[doi:10.11992/tis.201609016]
[5]齐小刚,胡秋秋,刘立芳.基于MapReduce的并行异常检测算法[J].智能系统学报,2019,14(2):224.[doi:10.11992/tis.201809007]
QI Xiaogang,HU Qiuqiu,LIU Lifang.Parallel anomaly algorithm based on MapReduce[J].CAAI Transactions on Intelligent Systems,2019,14():224.[doi:10.11992/tis.201809007]
[6]逄增治,郑修楠,李金屏.全钢子午线轮胎X光图像的缺陷检测研究现状[J].智能系统学报,2019,14(4):793.[doi:10.11992/tis.201806014]
PANG Zengzhi,ZHENG Xiunan,LI Jinping.Research status of defect detection in X-ray images of all-steel radial tires[J].CAAI Transactions on Intelligent Systems,2019,14():793.[doi:10.11992/tis.201806014]
[7]王春凯,庄福振,史忠植.易变数据流的系统资源配置方法[J].智能系统学报,2019,14(6):1278.[doi:10.11992/tis.201908011]
WANG Chunkai,ZHUANG Fuzhen,SHI Zhongzhi.System resource allocation for variable data streams[J].CAAI Transactions on Intelligent Systems,2019,14():1278.[doi:10.11992/tis.201908011]
[8]王凯旋,任福继,倪红军,等.面向电力设备红外图像的温度值识别算法[J].智能系统学报,2022,17(3):617.[doi:10.11992/tis.202105043]
WANG Kaixuan,REN Fuji,NI Hongjun,et al.Temperature value recognition algorithm for the infrared image of power equipment[J].CAAI Transactions on Intelligent Systems,2022,17():617.[doi:10.11992/tis.202105043]
[9]戚银城,耿劭锋,赵振兵,等.基于特征迁移的螺栓图像超分辨率处理方法[J].智能系统学报,2023,18(4):858.[doi:10.11992/tis.202201009]
QI Yincheng,GENG Shaofeng,ZHAO Zhenbing,et al.A method for super resolution processing of bolt image based on feature transfer[J].CAAI Transactions on Intelligent Systems,2023,18():858.[doi:10.11992/tis.202201009]
[10]高海洋,张明川,葛泉波,等.基于点集匹配的缺陷样本图像生成方法[J].智能系统学报,2023,18(5):1030.[doi:10.11992/tis.202209028]
GAO Haiyang,ZHANG Mingchuan,GE Quanbo,et al.Method of defect sample image generation based on point set matching[J].CAAI Transactions on Intelligent Systems,2023,18():1030.[doi:10.11992/tis.202209028]
备注/Memo
收稿日期:2024-2-22。
基金项目:安徽省重点研究与开发计划项目(2022107020030).
作者简介:蒋世杰，硕士研究生，主要研究方向为计算机视觉、异常检测。E-mail：jiangshijie@mail.ustc.edu.cn;夏秀山，副研究员，主要研究方向为数字图像处理、计算机视觉、人工智能。主持、参与国家和省部级科研项目10余项，发表学术论文10余篇。E-mail：xiaxiushan@iat.ustc.edu.cn;翟伟，博士后研究员，主要研究方向为计算机视觉、具身智能、多模态分析。主持国家自然科学基金青年项目1项，发表学术论文40余篇。E-mail：wzhai056@ustc.edu.cn。
通讯作者:夏秀山. E-mail：xiaxiushan@iat.ustc.edu.cn
更新日期/Last Update:
2025-03-05